Itinerary

Opening keynote. Every second between 'save' and 'signal' shapes what your engineers build. Yuki Tanaka on why the inner development loop deserves the same product rigor as anything you ship to customers — with before/after numbers from three companies that invested.

Our monorepo CI took 40 minutes on a good day. This talk walks through how we cut it to 6 minutes with content-addressed caching, remote execution, and a test-selection model — including the two migrations that failed first. You'll leave with a decision framework for which incremental-build investments pay off at which repo sizes, and the graphs to convince your platform team. This session now includes a live demo of remote build caching.

The five developer-productivity metrics that survived two years of pruning, the eight that got deleted, and how to report numbers without starting a stack-ranking arms race.

Canary deploys, feature flags, and automatic rollback for 300 services — run by two SREs. The boring architecture that makes it possible.

Shipping golden-path templates under regulatory constraints: what compliance actually required, what engineers actually adopted, and the three template decisions that drove 80% adoption.

Code generation is easy; trusting it is hard. This session covers verification patterns for AI-generated code — property tests, mutation coverage, snapshot judges, and CI gates — with data from 18 months of running them on a 200-engineer codebase. Includes what we stopped doing because it didn't catch anything.

Your compiler already knows which tests could break. CI throws that knowledge away. How to wire compiler incrementality facts into test selection.

Hands-on workshop: build a content-addressed remote build cache from scratch, break it with the classic invalidation bugs, and fix them. Bring a laptop with Docker and Go.

Big migrations fail socially before they fail technically. A sequencing playbook from six large-repo migrations — and the one dashboard that keeps executives from canceling at 60%.

Panel. Three people who have run both models at scale argue it out with numbers instead of vibes: build times, ownership boundaries, dependency drift, and what they would choose starting today.

Keynote. Internal platforms fail when they're run as cost centers and succeed when they're run as products. Amara Diallo on user research for engineers, roadmaps that say no, and the adoption curve of a platform people actually want.

A maintainer's view of running a critical open-source build tool: ecosystem testing before releases, what a deprecation policy costs, and what downstream users do that terrifies us.

New engineers decide how they feel about your platform in the first five minutes. Rebuilding onboarding around that window — current time-to-first-deploy record: 11 minutes.

A 10-minute tour of turning a static docs site into one that answers questions with citations, stays honest when it doesn't know, and costs under $50/month to run. Live demo, real failure cases, and a checklist you can apply to your own docs this week.

Lightning talk. The day our feature-flag file hit 4,000 lines, the outage that followed, and the three rules that fixed it.

Lightning talk. A rapid-fire demo reel of our AI code-review bot getting it wrong — flaky approvals, phantom refactors, confidently wrong comments — and the guardrail that fixed each one.

Spans on everything: ticket pickup, review wait, CI stages, deploy. An end-to-end trace of how work actually flows — and the three bottlenecks it exposed that no metric had.

A practical migration path from grep debugging to structured, queryable logs: schema conventions that survive refactors, sane sampling, and the queries that replaced our five favorite grep incantations.

Hands-on workshop: content-pinned toolchains, per-repo shells, and CI that runs byte-identical to laptops. From 'works on my machine' to one-command environments.

Two years, 60 self-hosted agents, one six-week parallel-run. The unvarnished story of a CI platform migration, with the cost curve before and after.

Provenance, signing, and dependency review shipped as defaults developers barely notice — and a break-glass path that gets audited instead of blocked.

Panel. Open Q&A with three platform leaders. Bring the question you can't ask at your own company — adoption fights, funding fights, and the projects they killed.

A year after instrumenting the developer-experience org at Substrate: which investments moved the numbers, which dashboards got deleted, and the survey question that predicts attrition better than any metric.

Closing keynote. Bianca Moretti on cutting release lead time from six weeks to daily deploys across an embedded-plus-cloud stack — and why the biggest wins came from deleting process, not adding it.