June 11 - 15, 2026
Web Engineering Summit
Amsterdam & Online

Web Engineering Summit 2026

The full-stack excellence conference

Web Engineering Summit is a 2‑day, 2‑track conference for engineers building and evolving production web systems. Explore how modern web platforms are architected, scaled, observed, secured, and improved over time — with a focus on reliability, performance, quality, and developer experience. Join thousands of engineers who care not just about shipping features, but about building systems that last.



Esta edición del evento ha finalizado, las últimas actualizaciones de este Tech Conference están disponibles en el sitio web de la marca.

GitNation Events Supporter

Catch, trace, and fix bugs across your entire stack Learn more

Agentes en el Canvas Con tldraw
29 min
Agentes en el Canvas Con tldraw
The Talk delves into the exploration of Agents on the Canvas, integrating LLMs with visual elements and spatial relations. It covers the development of the Agent Starter Kit for navigating Canvas tasks and introduces fairies as collaborative agents. The discussion includes managing tasks, utilizing Cloud Code for project collaboration, creating interactive UIs, and customizing UI components. The evolution of agentic canvas code, agent interactions, user actions, token usage, and data handling with LLMs are highlighted, emphasizing experimentation and optimization for SDK integration.
El Código Muerto No Debería Existir: Cómo Eliminamos 28k Líneas de Código, Un Knip a la Vez
35 min
El Código Muerto No Debería Existir: Cómo Eliminamos 28k Líneas de Código, Un Knip a la Vez
The Talk covers the importance of building a design system for efficient component usage and maintaining a clean codebase for faster shipping. It discusses the impact of unused code on development efficiency, challenges in cross-module usage analysis with TypeScript, and the use of KNIP for identifying and removing unused code elements. KNIP aids in code cleanup, optimization, and dependency removal, enhancing code cleanliness and efficiency. The Talk also addresses challenges in adapting to KNIP in CI workflows, caution in dynamic codebases, and the potential for KNIP to expand to other cleanup tasks beyond code size reduction.
Chunking
28 min
Chunking
Tobias Koppers from Vercel explains the process of chunking in bundlers and its impact on module organization and loading efficiency. Optimization strategies involve analyzing metrics, chunk groups, request counts, and long-term caching for better performance. Managing trade-offs in chunking with long-term caching, correctness, and CSS order considerations is crucial. Challenges include implicit CSS dependencies, overshipping, and balancing CSS ordering for future optimizations. Compile time complexities and future optimization challenges highlight the evolving nature of chunking strategies.
Pon a prueba tus reflejos (y mi aplicación)
32 min
Pon a prueba tus reflejos (y mi aplicación)
Eddie from PayPal introduces a Whack-A-Mole game to discuss real-time data and engage the audience. Real-time data is essential for dynamic updates and efficient data handling. Challenges with AI-generated MVP led to a successful times three model with real-time features. The evolution of data communication technologies from Comet to WebSockets is highlighted. Transition to WebSocket native implementation for efficiency and benefits of using Superbase are discussed. Real-time features in games, event handling, presence tracking, and event broadcasting are explored. React SDK, real-time tools, scalability with Meteor, and managing data noise are addressed in the Q&A sessions.
Bootcamp de Agente de Voz: Construye Agentes Listos para Producción desde Cero
130 min
Bootcamp de Agente de Voz: Construye Agentes Listos para Producción desde Cero
Workshop
Jesse Hall
Jesse Hall
¿Listo para pasar de cero a un agente de voz desplegado?La IA de voz avanza rápidamente, pero la mayoría de los tutoriales terminan en "hello world" y te dejan preguntándote cómo realmente llevar algo a producción. Este masterclass intensivo te lleva a través de todo el viaje: construir un agente de voz completo con TypeScript y React, desplegarlo en producción, configurar la observabilidad y crear marcos de evaluación, todas las piezas que necesitas para aplicaciones JavaScript del mundo real.Lo que construirás:Un agente de voz full-stack usando el SDK de TypeScript de LiveKitInterfaz de usuario de React con visualización de audio en tiempo real y gestión de estadoDespliegue en producción con la observabilidad adecuadaPruebas automatizadas y marcos de evaluaciónTemas cubiertos:Configuración de flujos de audio WebRTC con TypeScriptIntegración de modelos STT, LLM y TTS en tu stack de JavaScriptConstrucción de interfaz de usuario de React para interacciones de voz con estado en tiempo realManejo de interrupciones y turnos naturalesDespliegue en producción (Vercel)Configuración de observabilidad y reproducción de sesionesConstrucción de marcos de evaluación para aseguramiento de calidadEsto no es una demostración, es un bootcamp. Escribirás código TypeScript, depurarás problemas y desplegarás en producción. Al final, tendrás un agente de voz funcional y la confianza para construir más.
Construyendo Aplicaciones Fullstack Con Cursor
164 min
Construyendo Aplicaciones Fullstack Con Cursor
Workshop
Maurice de Beijer
Maurice de Beijer
Dirige el código. No solo lo escribas.El panorama del desarrollo está cambiando de la codificación manual a la orquestación de sistemas. Ya seas un ingeniero de React experimentado que busca multiplicar por 10 su producción o un creador orientado al producto que busca convertir un prototipo en una aplicación funcional, esta masterclass es tu puerta de entrada al desarrollo nativo de IA.Únete a nosotros en una masterclass práctica de 4 horas donde construiremos una aplicación fullstack rica en características desde cero usando Cursor Composer. Omitimos el código repetitivo y nos centramos en los flujos de trabajo exactos utilizados para enviar características de alta calidad en minutos, no días.Lo Que DominarásOrquestación de IA: Avanza más allá del autocompletado simple para gestionar cambios en múltiples archivos y lógica de aplicaciones complejas.El Marco .cursorrules: Aprende a imponer estándares de proyecto y prevenir "alucinaciones" de IA para un código consistente y profesional.Andamiaje Fullstack Rápido: Genera esquemas de base de datos seguros por tipo y lógica de backend al instante.Iteración de UI de Alta Fidelidad: Transforma indicaciones en interfaces pulidas y responsivas usando CSS moderno y bibliotecas de componentes.Endurecimiento y Refactorización: Usa IA para cazar errores, limpiar código "espagueti" y generar suites de pruebas completas.El ProyectoConstruiremos una aplicación web moderna y basada en datos que demuestre el poder del desarrollo asistido por IA. Al final de la sesión, tendrás una base de código funcional que muestra relaciones de datos complejas, UI dinámica y gestión de estado optimizada.
¿Para Quién Es Esto?Desarrolladores que quieren eliminar tareas repetitivas y centrarse en la arquitectura de alto nivel.Fundadores y Prototipadores que necesitan cerrar la brecha entre una idea y un producto funcional rápidamente.Entusiastas de la Tecnología ansiosos por dominar el kit de herramientas profesional nativo de IA.Requisito: Trae una laptop con Cursor instalado. Nosotros proporcionamos la hoja de ruta; la IA proporciona la velocidad; tú proporcionas la visión.
Como esta es una sesión en línea en vivo, te animamos a registrarte lo antes posible para recibir la guía de configuración previa a la masterclass y asegurarte de que estás listo para construir desde el primer minuto.
How I Taught LLMs How to Svelte
33 min
How I Taught LLMs How to Svelte
The talk by Paolo Ricciuti introduces Svelte framework basics, syntax evolution in Svelte 5, challenges with LLMs in Svelte, Model Context Protocol (MCP) for LLMs, utilizing MCP for code verification, enhancements with Svelte MCP CLI, Svelte benchmark creation, integration of agent skills in node modules, Svelte adoption and DX differentiation, and Svelte autofixer and MCP linter.
Plantillas y Componentes para Claude Code: El Futuro de los Flujos de Trabajo de Codificación de IA
30 min
Plantillas y Componentes para Claude Code: El Futuro de los Flujos de Trabajo de Codificación de IA
Daniel Avila discussing templates and components for Cloud Code, emphasizing the importance of configuration settings and skills integration in the development process. Cloud Code settings, commands, sub-agents, and skills integration play a crucial role in developing a complete template for AI agents. Skills injection, documentation, development template, Cloud Code components sharing, harness engineering, logs review, process improvement, configuration levels. Configuration levels, settings.json, hooks, verification, deterministic projects, templates, status lines, cloud.md, best practices. Explaining skill invocation control, model usage, and skill specialization for message crafting within commits. Emphasizing the importance of managing the context window effectively for skill injection and understanding the interaction with various agents and tools.
Sistemas Basados en Agentes Operativos - Visión General, Configurar, Ejecutar, Orquestar, Monitorear
129 min
Sistemas Basados en Agentes Operativos - Visión General, Configurar, Ejecutar, Orquestar, Monitorear
Workshop
Alex Korzhikov
Pavlik Kiselev
2 authors
Esta masterclass explora cómo operan los agentes independientes a nivel de tiempo de ejecución y cómo se diferencian de los pipelines de IA tradicionales. Examinaremos la arquitectura de los agentes, los bucles de planificación, los modelos de memoria y la ejecución de herramientas. También cubriremos la coordinación de múltiples agentes, incluyendo el aislamiento de estado y el control de recursos. Un enfoque clave es la seguridad y la gobernanza: acceso basado en capacidades, sandboxing y riesgos de inyección. Finalmente, abordaremos la observabilidad y la supervisión: rastreo del razonamiento, auditoría del uso de herramientas e implementación de mecanismos de control para sistemas de producción.

Todos los ejemplos y conceptos estarán basados en la pila de Node.js y exploraremos por qué Node.js es particularmente adecuado para construir tiempos de ejecución de agentes listos para producción, sirviendo como el plano de control para la supervisión, integración, ejecución en streaming y coordinación distribuida.

Referencias: OpenClaw, n8n, LangChain.
Agentes de IA Autónomos en Acción con el Método Ralph Wiggum
28 min
Agentes de IA Autónomos en Acción con el Método Ralph Wiggum
Today's Talk delves into the Ralf Wiggum method for autonomous AI agents, emphasizing persistent looping based on specifications in AI coding. It addresses context management challenges in AI chats, promotes optimized coding workflows with feedback loops and engineer decision-making. Automation of specifications and task implementation is highlighted, alongside interactive conversation, prototyping, and managing code implementation. The importance of code quality enhancement, exploring new approaches with AI coding agents, and ensuring smart application of AI tools is discussed. Utilizing impactful skills and convincing leadership with results from the Ralph method are key points, along with enhancing task efficiency and overcoming tool limitations for efficient work.
Caminando por el Camino Pavimentado de Netflix (Incluidos los Baches): Web Framework, Hawkins Design System
26 min
Caminando por el Camino Pavimentado de Netflix (Incluidos los Baches): Web Framework, Hawkins Design System
Misha, a UI engineer at Netflix, discusses challenges faced by the UI platform and product team. Platform tooling and migration benefits focus on boosting productivity and efficiency. Tool selection and platform transition involve analyzing and bundling tools for streamlined processes. Legacy setup and community libraries highlight the complexity of pre-UiPR setups. Phased migration approaches and component transitions emphasize reducing technical debt and collaboration. Leadership justification, team collaboration, and UI framework creation and regrets are also key points discussed.
No Servers, No Cloud, No Masters: Make P2P Apps
32 min
No Servers, No Cloud, No Masters: Make P2P Apps
Working on decentralized peer-to-peer technology without infrastructure, enabling JavaScript apps to be deployed peer-to-peer without servers or cloud. Refining technology for four years, partnership with Tether, launch of Keet app for peer-to-peer connection using hole punching algorithm. Built with PAIRS stack for robust, cost-effective technology. BEAR - minimal Node-like runtime for peer-to-peer. PeerRuntime module enables peer-to-peer updates and simple primitives. Dual-distribution model for app deployment with peer-to-peer updates and cost savings. React Native boilerplate for mobile apps with peer-to-peer updates. Predators infiltration risk in peer-to-peer. Local devices turning into peers for updates.
Uso del Desarrollo Dirigido por Especificaciones para Flujos de Trabajo de Producción
19 min
Uso del Desarrollo Dirigido por Especificaciones para Flujos de Trabajo de Producción
Eric Hanchett discusses spec-driven development, emphasizing structured code creation with specifications before writing any code. The importance of context in guiding large language models and the evolving nature of software development is highlighted. Kiro, an AI IDE coding assistant, offers a comprehensive solution for improved spectrum development processes. Kiro's focus on vibe and spec mode caters to the need for working on larger and more complex projects. Spec-driven development in Kiro is beneficial for in-depth features and structured project building. The discussion also includes property-based tests, MVP creation, and the Model Context Protocol (MCP) integration within the spec-driven development flow.
Construyendo un Motor de JavaScript en Rust: Lecciones de Boa
24 min
Construyendo un Motor de JavaScript en Rust: Lecciones de Boa
Jason Williams discussed building a JavaScript engine in Rust, focusing on challenges and achievements with Servo, including making Servo work on Windows. Boa was introduced as a JavaScript engine project with a blog on Mozilla Hacks. Boa's development stages were highlighted, starting from file reading and tokenization. The talk delved into token types, lexical declarations, parsing, and execution methods in Boa. Community involvement, testing emphasis, and conformance achievements were emphasized, with a notable collaboration on Temporal implementation reaching 100% test pass rate.
Enseñemos a Neo Kung-Fu: Construyendo una Habilidad Agente con Herramientas MCP
60 min
Enseñemos a Neo Kung-Fu: Construyendo una Habilidad Agente con Herramientas MCP
Workshop
Misha Kazakov
Misha Kazakov
Los asistentes de codificación de IA están evolucionando de simples autocompletados a agentes autónomos que pueden interactuar con sistemas externos. Pero, ¿cómo enseñas a un agente a seguir tus flujos de trabajo y usar las herramientas adecuadas en el momento correcto?En este masterclass práctico, descubrirás:Qué es MCP (Model Context Protocol) y cómo estandariza la integración de herramientas para LLMsCómo los Agent Skills empaquetan el conocimiento y los flujos de trabajo específicos del dominio para agentes de IALas diferencias clave entre servidores MCP y Skills, y cuándo usar cada unoCómo crear un Skill personalizado que orqueste múltiples herramientas MCP en un flujo de trabajo cohesivoAl final de esta sesión, construirás tu propio Skill de IA que utiliza herramientas MCP — una habilidad de lucha de Neo con temática de Matrix.Quién debería asistir: Este masterclass es ideal para desarrolladores de software que usan asistentes de codificación de IA (Claude Code, Cursor, o similares) y desean extender sus capacidades con integraciones y flujos de trabajo personalizados. No se requiere experiencia previa en MCP o Skills — solo trae tu curiosidad y una laptop.
Interfaces Agénticas: Herramientas, Habilidades, UI Generativa y Web MCP
31 min
Interfaces Agénticas: Herramientas, Habilidades, UI Generativa y Web MCP
Wes discusses agentic interfaces and the changing interaction with the web. Exploring different software interfaces from UI-less to tailored interfaces and interconnected apps. Implementation options include generative UI with basic, declarative, and fully generative levels. Exploring the challenges of integrating unconventional components like wacky shopping lists and the need for a balance between dynamic and fully generative components in user interfaces. Exploring the shift from fast interactions to slower MCP interfaces, emphasizing the need for adaptive Smart Home systems and user-centric interfaces. Excitement about WebMCP, enabling extended website functionality through WebSemantics and imperative JavaScript API. Utilizing agents for enhanced website control, switching between Cloud and browser for varied experiences. Impact of evolving technology on business adaptation and the shift towards platform control. Discussion on the future of dynamic UI with user-driven component rendering and the importance of preserving the openness of the web amidst AI-driven application generation. Contemplation on the effectiveness of MCP-based interactions as AI companies transition to charging for services, hinting at potential cost implications in the near future.
Monitoreo FullStack con Open Telemetry: Observabilidad de Extremo a Extremo para Aplicaciones Modernas
30 min
Monitoreo FullStack con Open Telemetry: Observabilidad de Extremo a Extremo para Aplicaciones Modernas
The speaker discusses the importance of performance development in AI applications, concerns about application crashes and vulnerabilities, utilizing data for performance improvement, application monitoring with OpenTelemetry, integrating AI for enhanced security measures, streamlining data analysis and reporting, leveraging AI for security threat detection, optimizing data logging, and involving data experts for determining data importance.
MemLab: Automatización de la Detección de Fugas de Memoria y Análisis de Heap
19 min
MemLab: Automatización de la Detección de Fugas de Memoria y Análisis de Heap
Liang introduces MemoryLab, a tool for automating JavaScript memory leak detection. MemoryLab aims to make memory debugging more accessible and scalable by automating the process and utilizing a three snapshot differentiation approach. It clusters similar leak traces to reduce debugging overload and presents representative cluster retainer traces for each memory leak cluster. MemoryLab integrates AI coding tools for memory detection, analysis, and debugging, with the MCP server enabling accurate investigation of memory leaks in heap snapshots.
Creando un Sistema de Diseño para 1B+ Usuarios en la Era de la IA
25 min
Creando un Sistema de Diseño para 1B+ Usuarios en la Era de la IA
Noe Yamamoto, a UI engineer, shares his journey from Squarespace to WhatsApp, highlighting a passion for UI. WhatsApp Web has evolved to a feature-rich platform with a consistent design language. Engineers face challenges in maintaining design consistency and feature parity with mobile. They also encounter issues in identifying design inconsistencies and usability problems. A UI refresh in 2024 brought significant design improvements, with positive user feedback. Legacy component management and AI integration present challenges, as does initiating and evolving the design system. Community-driven design systems and AI integration are emphasized, along with testing approaches and the importance of accessibility in design systems.
Código Avanzado de Claude — Flujos de Trabajo de Producción, Subagentes y Ejecución Autónoma
207 min
Código Avanzado de Claude — Flujos de Trabajo de Producción, Subagentes y Ejecución Autónoma
Workshop
Aleksei Petrov
Aleksei Petrov
La mayoría de los desarrolladores que utilizan herramientas de codificación de IA se topan con el mismo muro en proyectos reales: las demostraciones parecen mágicas, pero el trabajo de producción se desmorona. Este masterclass muestra por qué y qué hacer al respecto.La lección principal: la ejecución es rápida y barata solo cuando la preparación es profunda.Lo que construiremos, en vivoDesde un directorio vacío hasta una aplicación de producción desplegada: una aplicación de conferencia móvil con el horario del evento, perfiles de oradores, búsqueda y favoritos. Los asistentes la abrirán en sus teléfonos antes de que termine la sesión.
Las cuatro etapas de envío con IAVerás un flujo de trabajo de producción completo, deliberadamente orientado hacia el trabajo que realmente determina la calidad:- Investigación — Reuniendo todo lo que la IA necesitará por adelantado: documentos oficiales, patrones de código, datos reales. Para el momento de la ejecución, no hay que buscar nada. Aquí es donde proviene la calidad.- Planificación — Dividiendo la aplicación en tareas claras con criterios de aceptación, escritos en vivo. Saltarse esto es la razón por la que la mayoría de las construcciones de IA fallan.- Ejecución — Diez tareas ejecutándose en paralelo a través de git worktrees, orquestadas por Navigator (un plugin de Claude Code) en una pila de React/Next.js. Código y documentos escritos juntos.- Revisión y Envío — Puertas de calidad (pruebas, lint, tipos, construcción), commits limpios, fusión, despliegue en Vercel. URL en vivo compartida con la sala.
Bonus: Humano + Claude Code vs. agente completamente autónomoMientras la construcción en vivo se ejecuta, Pilot — nuestro agente de codificación autónomo — construye la misma aplicación a partir de la misma especificación, por su cuenta, en un repositorio separado. Al final, abrimos ambos lado a lado: dos aplicaciones funcionales, dos historiales de pull request, dos URLs en vivo. Una impulsada por humanos, otra completamente autónoma.
Te irás conUn flujo de trabajo repetible para enviar software real con Claude Code, además de técnicas concretas que puedes aplicar a tus propios proyectos el lunes por la mañana.
Open Source Voice AI: Cómo Construimos la Infraestructura de Modo de Voz de ChatGPT
20 min
Open Source Voice AI: Cómo Construimos la Infraestructura de Modo de Voz de ChatGPT
Voice interaction is evolving rapidly with AI agents moving beyond chat windows. Achieving natural conversations with voice AI presents significant challenges, especially in meeting real-time interaction expectations. Advancements in real-time conversation technology and infrastructure are improving latency gaps in voice apps. Optimizing latency at each stage of voice agent processing is crucial for efficient development. Semantic turn detection and efficient function calling are key to optimizing user experience in Voice AI. Workflow management, cost management, signal recognition, and observability are vital aspects in voice AI development.
AI Puede Generar Pruebas para Ti, ¿Y Ahora Qué?
22 min
AI Puede Generar Pruebas para Ti, ¿Y Ahora Qué?
Living in an exciting and productive period with AI's assistance in test generation and scalability, led by Maya, a Staff Software Engineer passionate about AI accessibility. Streamlined test generation through AI can compromise code quality if not placed correctly, emphasizing the importance of test validity over quantity. Challenges of test generation include cost impacts on code quality, highlighting the need for validity over quantity and the importance of placing tests correctly. Understanding risks and validation criteria in code testing, with an example focusing on search functionality. Importance of mapping tests to specific quarens for effective validation in Agile Testing. Exploring AI integration into Agile Testing, balancing AI and human input for effective workflows. Transitioning to a new scalable workflow with AI test plan generation for enhanced efficiency. Establishing collaboration standards with AI, prioritizing strategic testing tactics over quantity and ensuring code safety beyond complete coverage.
Cómo Usamos AI para Construir TanStack AI
33 min
Cómo Usamos AI para Construir TanStack AI
Zalim Tuzlak explains features of building 10 stack AI, enhancing capabilities with middleware and tool code mode, efficient AI tool script writing, challenges and breakthroughs in AI development, building scalable AI tools with activity-based tree shaking, early AI concepts and code quality strategies, empowering rapid AI iteration, TensTech AI beta release with Angular support, AI model comparison and evaluation, AI SDK features enhancement, quality assurance processes for TanStack AI docs, personal stress and decision fatigue in AI development, challenges in middleware development with AI assistance, skill creation process within the team, and TanStack AI's end goal and selling point.
HTML en Canvas: Uniendo UI y GPU en la Web
17 min
HTML en Canvas: Uniendo UI y GPU en la Web
Santiago Colombatto
Tomas Ferreras
2 authors
Santiago Colombato discusses challenges in synchronizing DOM and WebGL for visually heavy websites, focusing on scroll sync and basics of WebGL's rendering process. The talk addresses the complexities of rendering text on GPU in WebGL and the challenges of HTML in canvas implementation. The proposal introduces advanced features like a paint event and element image capture to enable synchronization of HTML changes and element snapshots. Exciting prospects include blurring the lines between DOM and canvas, showcasing 3D scenes, distorted text, and GPU effects, offering new possibilities for website generation.
Supercharging Your Tooling With Rust
24 min
Supercharging Your Tooling With Rust
Mael discusses the evolution of Yarn from its origins in 2016 to the current version being developed in Rust. Yarn's stability and familiarity with the codebase led to the decision to build Yarn 6 in Rust, despite challenges faced with previous rewrites and limitations of Yarn 4. Large companies using Yarn faced latency issues due to the number of workspaces. Yarn aimed for improved performance with features like lazy install to enhance user experience. Yarn aimed to be exceptionally fast, catering to various ecosystems beyond JavaScript, like Python and Rust, to enhance usability. Yarn faced challenges with breaking changes and friction during the rewrite to Yarn 2, emphasizing the importance of careful implementation in Yarn 6's development. Learning a new language was a key part of the rewrite process, with significant progress in test passing and personal proficiency in Rust. Developing in Rust led to a growth in personal proficiency and revealed insights into Rust's tooling and concurrency benefits. Rust revealed experimental features and concurrency benefits, requiring mental model shifts. Rust's human-written code rewrite with AI assistance. Challenges of integrating Yarn rewrite with Rust version and freezing main branch for rewrites. Considerations for rewriting projects in Rust, YARN's improved performance, and the importance of motivation and resilience in such endeavors. YARN's expansion to support multiple ecosystems and the invitation to engage in devlogs and share feedback on the talk and rewrite.
Orquestando Flujos de Trabajo de Contenido a Escala de Netflix
20 min
Orquestando Flujos de Trabajo de Contenido a Escala de Netflix
Pradeep Shah, senior software engineer at Netflix, discusses the coordination of rule engines, ML models, and humans in content understanding infrastructure. Building an extensible system based on a shared framework derived from past workflow challenges. Implementing a two-tier architecture for state management and workflow execution using Conductor. Utilizing Conductor for orchestration with retries, timeouts, and parallel execution. Workflow onboarding steps include defining triggering events, task creation, lifecycle tracking, and human intervention for AI confidence below threshold. Human decision structured output, model feedback pipeline, and domain-agnostic schemas for reusability. AI agent service mirrors human process, seamless AI integration, and architecture success factors: separation of concerns, pluggability, multi-modal workers, and human corrections as training signals.
Migración de Bibliotecas de Componentes de UI Potenciada por LLM
25 min
Migración de Bibliotecas de Componentes de UI Potenciada por LLM
Nawal discusses using LLMs for migrating component libraries at Zalando, highlighting challenges and complexities. Discussing a recent migration problem at Zalando involving unification of component libraries and challenges in automating migration process. Encountering limitations with codemods, leading to exploration of LLMs for automated migration processes. Conducting iterative experiments to enhance LLM accuracy in component migration. Building a CLI tool for component migration. Using AIDAR for automating git commits & maintaining LLM accuracy. Recommendation for team closeness to code, successful complex transformations by LLM, challenges with hallucination and ignoring instructions. Lessons learned include field addition, non-deterministic outputs, slow responses, project constraints, and tool evaluation. Evaluating tool enhancements, PROMPT best practices, core approach, improved tools, LLMs for simple migrations, and tool limitations with multiple file changes.
Click. Ship. Done. AI Agents on Cloudflare
22 min
Click. Ship. Done. AI Agents on Cloudflare
AI evolution from the past to present, showcasing advancements like ChatGPT for easier programming accessibility. Generative AI advancements with ChatGPT, followed by rapid evolution in coding accessibility and introduction of innovative AI models like Zorro by OpenAI. User input processed by LLM in chatbot scenarios. Agentic loop adds steps for evolution and manipulation in processing prompts. LLM guides steps in user interaction. MCP's role in LLM operations and potential challenges of self-hosting agents. Future uncertainty in self-hosting with Cloudflare workers and utilizing durable objects for state storage in APIs. Interactivity improvement with durable object storage in Cloudflare workers and the integration of AI chat agents for message handling. Accessing LLM models, utilizing screen text for message handling, integrating MCP tools, and deploying AI chat agents on Cloudflare. Deploying AI chatbots on Cloudflare, connecting to MCP servers, and handling prompts from GitHub MCP. Configuring MCP server tools, deploying with wrangler, and exploring Cloudflare agent setup. Setting up Cloudflare chat agent, deploying with wrangler, leveraging Cloudflare tools for efficient deployment. Setting up Cloudflare chat agent with efficient deployment using wrangler deploy command and GitHub integration. Easily interact with AI chat agent and other functionalities on Cloudflare.
Piensa Como un Tester: Qué Buscar en el Código Generado por AI
16 min
Piensa Como un Tester: Qué Buscar en el Código Generado por AI
Developers write code to make things work. Wendy, QA Automation Engineer, introduces John, a great developer who encounters challenges implementing a new feature. Checkout is critical for the company's existence, emphasizing its importance. When a coupon expires, issues arise with discounts. John faces challenges fixing the code, revealing different perspectives between developers and testers. Developers and testers are partners balancing each other's perspectives. Introducing the QAI for adding QA insights to development processes. Ensuring code matches agreements is crucial. Phase B scans code for potential issues. Code alignment is the gate to address discrepancies between agreements and code. Product oversight in code scenarios. Code failures can be concealed and dangerous. Handling hidden failures is crucial in code development. User faced invisible failures; handled vs. hidden code distinction. Fast code requires a security lens and skepticism. Trust gap between front end assumptions and back end validations. Distinguishing code gaps; AI lacks awareness; critical accessibility issues; distinct code failure types with separate focuses. Utilizing team patterns for efficiency; AI effectiveness through direction with QAI; improving code quality with QAI assistance. Contradiction between PRD and Figma resolved by John; Efficient bug fixing and code scanning; Enjoying exploring edge cases and unique challenges. Importance of clean code for efficient development; Addressing bugs at the right time saves resources; Transition from bug finding in development to production.
La vida de un ESM en Node.js – y cómo está cambiando para mejor
29 min
La vida de un ESM en Node.js – y cómo está cambiando para mejor
Talk covers the life of an ESM in Node, import resolution, module loading improvements, TypeScript transpilation challenges, customization of module resolution and source loading, handling compilation and dependencies, ESM customization, ESM limitations, circular dependencies, and optimizing loading performance in Node.
MCP Apps – el Próximo Web
33 min
MCP Apps – el Próximo Web
Liad Yosef
Ido Salomon
2 authors
The Talk discusses the evolution of MCP-UI, its partnership with OpenAI and Anthropic, and its adoption by early adopters like Shopify. It emphasizes the community-driven maintenance and support of MCP-Apps, transmitting UI over MCP, and standardizing app communication with hosts. The exploration of interactive UI components, real-time model interactions, and the revolutionization of web experiences with UI chunks are highlighted. The Talk also delves into enhancing user experiences with adaptive apps, future directions for MCP apps, global reach, security measures, developer control, user engagement, and integration of MCP apps with WebMCP.
El Estado de AI para el Desarrollo Web
35 min
El Estado de AI para el Desarrollo Web
Sasha discusses AI for web development, based on web dev AI survey results. AI's pervasive presence highlighted through personal experiences and industry trends. Increase in AI usage reflected in code generation and refactoring practices, impacting workflows and costs. Top pain points of AI tool usage include hallucinations, inaccuracies, and rising financial costs. Concerns about AI impact on job security, gender divide in environmental impact perception, and Moore's Law implications on rising AI costs. Speculation on faster, cheaper AI future, self-evolving code, and digital immune system. Software evolution towards probabilistic systems, personalized codebases, and digital DNA. AI potentially controlling digital lives, rendering software tools obsolete, and prompting critical thinking on future implications. Insights from AI survey results showcasing AI usage trends based on experience levels. Surprising impact of AI usage and philosophical views on AI terminology. Implications of AI advancement on coding skills, adaptation of software skills for the AI era, and concerns over AI-generated tech debt and social impacts. Speculation on AI tokens as global currency, ethical considerations, and impact on coding skills. Contemplation of AI future consequences and preparation for potential outcomes.
Halving Your CI Pipeline – Estrategias Prácticas de Optimización
10 min
Halving Your CI Pipeline – Estrategias Prácticas de Optimización
Serhii discusses optimizing CI pipelines, reducing pipeline time by 64% through structured approaches. Challenges in CI health, high failure rates, and the importance of trust in the system are highlighted. Prioritizing impactful optimizations, instance migration choice based on variance impact, and optimizing test environment sizing for efficiency. Reclaiming resources, optimizing CI processes, and enhancing stability through strategic changes. Building an advanced test retry system to enhance CI health and save costs. Key takeaways: Measure before optimizing, prioritize by return, optimize the whole system for improved efficiency and coverage.
¡Deja de Usar JSON Web Tokens (JWTs) para Autorización!
12 min
¡Deja de Usar JSON Web Tokens (JWTs) para Autorización!
Sohan Maheshwar humorously recounts feedback on advising against JWTs for authorization, highlighting challenges in payload structure and fine-grained permissions. Scalability and token revocation present issues, along with predicting permissions. Risks of overprovisioning permissions are addressed, with Macaroons as a solution and complexity in access control systems. Modern authorization systems like Zanzibar and Policy Engine prioritize correctness, flexibility, and security over simplicity of JWTs.
De 8 años a 6 meses: Orquestando agentes de IA para la refactorización a gran escala
29 min
De 8 años a 6 meses: Orquestando agentes de IA para la refactorización a gran escala
Alon Segal
Amit Hanoch
2 authors
We see how Monday used AI to break down a decade-old client-side monolith into micro-frontends, ensuring a complex software with long sessions and open tabs remained functional. The AI tool, Morphix, autonomously split prompts and validated tasks for consistent execution. The team focused on safe code migration, automated PR merging, and collaboration between AI and humans. They also optimized PR reviews, leveraged MorphX for monorepo restructuring, and managed AI costs effectively during the extraction process.
Auth Bajo Ataque Detectando Fallos de Auth de JavaScript en Producción Rápidamente
21 min
Auth Bajo Ataque Detectando Fallos de Auth de JavaScript en Producción Rápidamente
Viola, senior software engineer in London, presents Auth Under Attack to catch JavaScript auth failures in production fast. Auth failures lead to trust state disagreements. Tracing trust boundaries is crucial for accurate investigations. Understanding common failure shapes aids in recognition and resolution. Ensuring strong recovery flows and controls for trust extension are crucial. Permission leaks can occur when old authority persists. Focusing on system inconsistencies and broken trust chains is crucial for diagnosis and resolution. Mapping the system before implementation accelerates issue detection. Groundpact aids in enforcing authentication contracts pre-production, enhancing system security. Authentication aims for consistency and predictability, with CI as a crucial checkpoint to prevent incidents.
Una Breve Historia de Code Review (Y Qué Sigue)
8 min
Una Breve Historia de Code Review (Y Qué Sigue)
The speaker delves into the history of code review, highlighting the Fagan inspection process from the 1990s. Code review processes evolved from print-based reviews in the 1990s to collaborative online reviews via emails and diff tools in the 2000s. The modernization of code review processes includes the shift to pull requests with GitHub, challenges of increased PR reviews in the remote work era, and the role of AI in providing instant feedback on code quality. The future of AI-augmented code review envisions an asynchronous process with AI agents offering feedback through various channels, including CLI usage, and CodeRabbit's innovative approach to code review tools.
Construyendo Aplicaciones de ChatGPT y MCP con Toda la Comodidad del Moderno TypeScript DevX
21 min
Construyendo Aplicaciones de ChatGPT y MCP con Toda la Comodidad del Moderno TypeScript DevX
Fred, CTO of ALPIC, discusses MCP hosting, Skybridge framework, MCP evolution, and partnerships with major clients like OpenAI and Google. ALPIC has developed apps for the growing ChildGPT and Cloud Connector ecosystems. Skybridge, an open-source framework, facilitates easy app development. Explore various applications like the cottage app in the chatGPT ecosystem. Build and deploy apps using Skybridge and interact with the MCP server efficiently. Develop tools to view detailed information and deploy ChatGPT apps with Alpik Tunnel for easy access and testing.
400 Tech Leads. Mismos Problemas. Ninguno de Ellos Técnico
17 min
400 Tech Leads. Mismos Problemas. Ninguno de Ellos Técnico
Henri Fischer discusses tech lead challenges, emphasizing the importance of defining the tech lead role and addressing time management issues through delegation. Transitioning from a driver to a support role in one-on-one meetings improves efficiency and accountability. Strategies for effective feedback, delegation, and process management include reducing fear, setting clear expectations, and ensuring clear ownership for successful project management.
Escalando Monorepos de JavaScript a Nivel Empresarial: Lecciones de Más de 200 Paquetes
22 min
Escalando Monorepos de JavaScript a Nivel Empresarial: Lecciones de Más de 200 Paquetes
In the Talk, the most interesting ideas revolve around improving CI speed for over 200 monorepos by addressing waste in the CI process. Strategies include streamlining compilation processes, optimizing deployment by compiling only affected applications, and configuring self-hosted runners with S3 cache to enhance CI pipeline efficiency. Furthermore, enhancements in cache implementation, micro frontend build efficiency, and decoupling translations from CI to parallelize tests and reduce build and test times stand out as key points discussed.
Tomando un Dump: Usando Heap Dumps para Encontrar y Solucionar Problemas de Memoria y CPU en NodeJS
22 min
Tomando un Dump: Usando Heap Dumps para Encontrar y Solucionar Problemas de Memoria y CPU en NodeJS
Dan Shapir introduces heap dumps to identify and fix Node.js memory and CPU issues. Memory leaks in JavaScript occur when objects are still reachable, impacting memory usage. Continuous memory rise from leaks can deplete memory and lead to CPU usage escalation. Detecting and resolving memory leaks promptly is crucial to prevent system unresponsiveness in production environments. Heap dumps and memory dumps help analyze memory issues, with Chrome DevTools aiding in finding memory leaks. Pinpointing memory leaks involves comparing heap snapshots and analyzing leaked objects. Alternative methods for generating heap snapshots include new experimental node flags, automatic triggers, and built-in APIs, offering benefits but posing risks such as potential crashes and security vulnerabilities.
¿Reescribir o Refactorizar? Cómo Mover de Forma Segura Aplicaciones Legadas a Frameworks Modernos
21 min
¿Reescribir o Refactorizar? Cómo Mover de Forma Segura Aplicaciones Legadas a Frameworks Modernos
Sylwia Laskowska discusses challenges of frontend legacy migration, importance of legacy system migration for security and performance, considerations for big bang migration strategy, success story of Angular migration, incremental migration with strangler pattern, real-life migration story using feature flags, pitfalls in migration like prolonged timeline, and migration planning and execution strategies likened to renovating an old house.
Out-of-Order Streaming – El Futuro del Desarrollo Web
26 min
Out-of-Order Streaming – El Futuro del Desarrollo Web
Exploring web rendering evolution, limitations of server-side rendering, shift towards client-side rendering with JavaScript injection, evolution towards single-page applications with dynamic HTML generation, transition to server-side rendering, optimizing server-side rendering through hydration, partial hydration for efficient dynamic component rendering, server-side component rendering with external API data fetching, enhancing server-side rendering with streaming for improved response time, exploring sequential rendering for order control, implementing loading states with suspense and out-of-order streaming, integrating out-of-order streaming for efficient content replacement, optimizing streaming and browser functionality for seamless transitions and richer loading experiences.
De-bloating the Web: La Iniciativa de "Ecosystem Performance" (e18e)
20 min
De-bloating the Web: La Iniciativa de "Ecosystem Performance" (e18e)
The E18E initiative aims to modernize the JavaScript ecosystem by addressing redundant, bloated, and unmaintained packages. It focuses on reducing dependencies, upgrading packages, and improving project performance. Collaborators have successfully reduced package sizes and enhanced tools like Storybook and Prettier CLI. Tools like TinyGlobby and E18E CLI streamline dependency management for better project efficiency. Security enhancements, frontend alternatives, and collaborations with industry players are key aspects of the initiative's development. Future plans include tool enhancement, data visibility, and ongoing collaborations.
Lo Último de Deno
29 min
Lo Último de Deno
Leo from Dino talks about building desktop apps with Deno, transitioning web apps effortlessly, converting with various frameworks, and customizing easily. Experience app building with rdino desktop, live modifications with HMR flag, and the concept of transforming web apps into desktop apps. Deno offers bindings for efficient data transfer, lightweight backend options, and unified debugging. Discusses distribution methods, auto-updater with binary diffs, permissions, and sandboxing. Demo of a matrix client app, native UI integration, backend limitations, PWA support, rendering engine selection, and CSS features in Deno Desktop App.
Browser, API y Tecnología Asistiva: Un Triángulo Amoroso
28 min
Browser, API y Tecnología Asistiva: Un Triángulo Amoroso
Speaker at JS Nation in Amsterdam excited about the accessibility journey discussion. Emphasizes the research journey and accessibility relationship between companies and desktop journey. Importance of understanding browser's role in accessibility testing. Browsers create accessibility trees based on W3C guidance. Overview of platform APIs like AX API, UI automation, and iAccessible2. Complexities in browser, screen reader, and platform API combinations. Challenges with Chrome and Firefox interactions. Significance of bug reporting and centralized testing platforms for accessibility. ARIA attribute challenges and overlabeling issues. Promoting accessibility as a priority in companies. Initiating accessibility programs and promoting awareness through empathy labs. Using engaging conversations and empathy-centered approach for accessibility improvement.
Diseñando una Migración a Micro-Frontends
30 min
Diseñando una Migración a Micro-Frontends
In the last decade, Microfront Ends have seen widespread adoption in major companies, streamlining application development. Understanding Microfront Ends involves focusing on independence, fast flow, and migration goals over specific tools. Common architectures include application shell or UI composer and utilizing CDN for static files. Efficient deployment strategies include canary releases, edge computing, and rapid deployment frequency. Data sharing techniques include query strings, cookies, and event emitters for inter-microfrontend communication. Nurturing software growth and managing strategies for shared domain models and vendors are crucial for effective collaboration and development.
Depuración de Rendimiento Con AI
6 min
Depuración de Rendimiento Con AI
The speaker discusses the challenges of debugging rendering performance and the potential of using AI to address this issue efficiently. Emphasizes the limitations of current tools in providing detailed insights into rendering performance compared to JavaScript performance. Highlights the laborious and repetitive nature of manual optimization processes in improving rendering speed and the role of AI in automating such tasks. Discussing the importance of AI in simplifying complex tasks and flattening the learning curve, particularly in debugging performance issues. Highlighting the significance of installing Chrome developer tools MCP for effective AI utilization. Introducing techniques such as regression bisection and standalone reproduction for efficient performance issue identification and resolution. Exploring synthetic stress tests as a technique to magnify performance issues for better visibility. Demonstrating the power of creating extreme scenarios, like a page with 100,000 divs, to identify and replicate complex problems. Emphasizing the extensive capabilities of AI in systematically troubleshooting and proactively identifying performance bottlenecks.
Software Craftless: Escribiendo Código Que Haría Vomitar a una Cabra
15 min
Software Craftless: Escribiendo Código Que Haría Vomitar a una Cabra
The talk delves into the concept of writing complex code to boost creativity and appear smarter, using JavaScript challenges to highlight the importance of front-end development. It suggests adding confusion to code, utilizing complex functions, and mixing library versions. The future of front-end development involves emoji CSS, diverse HTML tags, and functional CSS. Architectural design principles include over-engineering systems, using diagrams with circles and hexagons, and organizing code by type. Effective testing strategies focus on code coverage, unit testing, and test order. Optimizing UI design, AI usage, and ethical AI utilization are essential, alongside audience engagement for feedback and social media interaction.
Una Fuga en el Shell – Cómo la Refactorización del Autocomplete Nos Rompió y Cómo lo Arreglamos
26 min
Una Fuga en el Shell – Cómo la Refactorización del Autocomplete Nos Rompió y Cómo lo Arreglamos
Glad to talk about JavaScript and Mongosh, a database CLI for MongoDB. Wanted to enhance autocomplete. Challenges with out-of-memory errors during testing. Debugging involved reducing heap size and using Node.js flags for memory analysis. Understanding memory retention challenges in JavaScript. Fixing memory retention bugs in Node.js core REPL for improved memory management. Introducing Finalization Registry and V8.QueryObjects for observing object finalization in JavaScript. Utilizing HeapSnapshots for memory leak debugging and AI integration. Challenges in understanding and debugging applications. Gap in testing autocomplete release with feature flag variations. Reliable ways to flush GC before heap inspection using v8.get heap snapshot and Node.js flag.
Black Friday: ¿Elegirías la Prueba de Rendimiento Correcta?
16 min
Black Friday: ¿Elegirías la Prueba de Rendimiento Correcta?
The talk discusses system failures in production despite passing tests, emphasizing the importance of spike testing during events like Black Friday. It highlights the consequences of not being prepared for sudden traffic spikes, such as errors, user dissatisfaction, and revenue loss. Common challenges in spike testing include thread pool saturation, database locks, and auto-scaling delays. Setting up a spike testing model involves identifying peak hours, establishing baselines, and creating realistic traffic distribution models to simulate scenarios like Black Friday.
Rustificando Vite: Diseñando una Cadena de Herramientas Híbrida para el Mundo Real
6 min
Rustificando Vite: Diseñando una Cadena de Herramientas Híbrida para el Mundo Real
The Talk explores JavaScript, TypeScript, Vite, and the move towards rolldown as a unified bundler. Rolldown combines the best features from previous tools, addressing challenges of multiple bundlers and complexity. It offers advanced chunking and a built-in minifier for optimized bundles. Integration of Rust-based bundles with JavaScript ecosystem through Swangular, filter functions, native Rust plug-ins, module resolution, TypeScript JSX transform, and toolchain V+ is discussed.
From Legacy to Delight: The Future of Node.js DX
21 min
From Legacy to Delight: The Future of Node.js DX
Claudio Bonder discusses the importance of Node.js documentation and the recent improvements made in developer experience. The evolution of Node.js API documentation has influenced tools and millions of users, with previous outdated tools leading to challenges in maintenance. The introduction of Dockit aimed to revolutionize Node.js API documentation with modern technologies. Dockit enhances documentation generation by supporting various sources, customization through AST and GSX transformation, and ensuring high-quality documentation for Node.js and community projects. The tool's innovative approach prioritizes user experience, collaboration, and strict common mark specification for efficient API doc creation.