Halving Your CI Pipeline – Practical Optimisation Strategies

JS Nation

JSNation 2026

Serhii Yakovenko

AI-Powered Code Review

TechLead Conference

TechLead Conf Amsterdam 2026: Adopting AI in Orgs Edition

Web Engineering Summit

Web Engineering Summit 2026

Every minute added to your CI pipeline costs real money and real developer time. At ~45 MRs per day and ~340 pipeline runs, a single minute of pipeline overhead translates to nearly 6 hours of lost engineering time daily. This talk presents the complete playbook we used to reduce our Merge Train pipeline from about an hour to about 22 minutes — a ~64% reduction — while improving CI health from the low 80s to the low 90s percent. Covering compute instance migration (with real benchmark data from 10 runs per configuration), service test extraction, CloudWatch optimisation, linter caching, flaky test detection, and cost analysis, attendees leave with a prioritised framework for their own CI optimisation efforts. Every optimisation includes before-and-after metrics, cost impact, and the trade-offs we navigated.

Every engineering organisation is experimenting with AI coding assistants, but few have built production-grade LLM integrations into their core developer infrastructure. I have such an experience, and I will share real patterns from deploying an AI-powered code review system across a 400+ person engineering organisation (~200 developers) — covering a competitive evaluation of 4 tools across 18 dimensions, building a webhook-based review architecture with slash commands and auto-review, evolving context enrichment from static rules to AI-powered document selection, managing a 4-model fallback chain on Vertex AI, and measuring impact through a feedback dashboard. Attendees will leave with a battle-tested playbook for integrating LLMs into their own engineering workflows — not as toys but as production infrastructure. Workshop Structure 1. The Code Review Bottleneck at Scale 2. Tool Evaluation — 4 Candidates, 18 Dimensions 3. Architecture — Webhook Server &amp; Auto-Review 4. Context Enrichment — From Path Rules to AI- Document Selection 5. Model Strategy — Migration &amp; Fallback Chain 6. Measuring Impact — Feedback Dashboard