Sessions
Wednesday 12 May
Elliot Ekström · Software Engineer, Eastbrook Digital
Our on-call rotation used to mean waking up to a wall of alerts with no idea where to start. This talk breaks down the incident command structure we adopted, th… Show more
Our on-call rotation used to mean waking up to a wall of alerts with no idea where to start. This talk breaks down the incident command structure we adopted, the severity rubric that keeps pages meaningful, and the postmortem format that actually changes behavior instead of just filing a report. You'll leave with a runbook skeleton and a blameless review template ready to adapt.
Frankie Lindqvist · Software Engineer, Fernway TechnologiesElliot Ekström · Software Engineer, Eastbrook Digital
Backward compatibility sounds simple until your third breaking change in a year quietly loses a partner integration. This talk covers a versioning strategy, a d… Show more
Backward compatibility sounds simple until your third breaking change in a year quietly loses a partner integration. This talk covers a versioning strategy, a deprecation timeline your consumers can actually plan around, and the contract tests that catch drift before it ships. You'll leave with a checklist for reviewing API changes before they leave code review.
Wren Whitcombe · Software Engineer, Cartwheel Software
Measuring developer productivity with lines of code or commit counts tells you who's gaming the metric, not who's effective. This talk covers the DORA-aligned m… Show more
Measuring developer productivity with lines of code or commit counts tells you who's gaming the metric, not who's effective. This talk covers the DORA-aligned metrics we adopted instead, how we avoided turning them into a surveillance tool, and the conversations they actually enabled with engineering leadership. You'll leave with a metrics dashboard your team won't resent.
Xan Chen · Software Engineer, Driftline
A monorepo without the right tooling turns every pull request into a full rebuild of everything, which nobody has patience for. This talk covers the dependency … Show more
A monorepo without the right tooling turns every pull request into a full rebuild of everything, which nobody has patience for. This talk covers the dependency graph and affected-target tooling we built to keep CI fast as the repo grew, and the ownership model that kept ten teams from stepping on each other. You'll leave with a checklist for evaluating whether your repo is ready to go mono.
Gale Sandoval · Software Engineer, Greenlight Data
A test suite that takes forty minutes to run is a test suite nobody trusts, so people start skipping it. We walk through how we diagnosed flaky and redundant te… Show more
A test suite that takes forty minutes to run is a test suite nobody trusts, so people start skipping it. We walk through how we diagnosed flaky and redundant tests, split the suite by risk tier, and got meaningful signal back in under five minutes. You'll leave with a triage framework for deciding which tests earn a spot in your fast path.
Harper Zabala · Software Engineer, Harborline Cloud
Dashboards full of metrics don't help when the one that would explain the outage isn't one anyone thought to track. This talk walks through moving from metrics-… Show more
Dashboards full of metrics don't help when the one that would explain the outage isn't one anyone thought to track. This talk walks through moving from metrics-first to trace-first debugging, the handful of high-cardinality tags that actually mattered, and the alert-fatigue cleanup that followed. You'll leave with a starter list of the traces worth instrumenting first.
Thursday 13 May
Devon Yilmaz · Software Engineer, Driftline
A new hire's first two weeks predicts how long they stay, and most onboarding docs are stale by the time anyone reads them. We cover how we rebuilt onboarding a… Show more
A new hire's first two weeks predicts how long they stay, and most onboarding docs are stale by the time anyone reads them. We cover how we rebuilt onboarding around a guided first ticket, a living runbook, and a buddy system with actual accountability. You'll leave with a 30/60/90 template you can adapt without a six-month rewrite.
Alex Delgado · Software Engineer, Northwind Systems
Feature flags start as a simple on/off switch and quietly turn into an unmanaged second database of business logic. This talk covers a rollout strategy that sep… Show more
Feature flags start as a simple on/off switch and quietly turn into an unmanaged second database of business logic. This talk covers a rollout strategy that separates release flags from experiment flags, plus the cleanup cadence that keeps stale flags from piling up. You'll leave with a checklist for auditing your flag inventory before it audits you.
Toni Brightwell · Software Engineer, Junction Point
Ad hoc releases meant every feature shipped whenever an engineer felt brave enough, which made planning impossible. This talk covers how we moved to a fixed-cad… Show more
Ad hoc releases meant every feature shipped whenever an engineer felt brave enough, which made planning impossible. This talk covers how we moved to a fixed-cadence release train, the feature-flag discipline that decoupled deploy from release, and the exception process for the rare hotfix. You'll leave with a cadence model you can adapt to your own team's risk tolerance.
Friday 14 May
Val Pereira · Software Engineer, Bluepeak Labs
You don't actually know your system is resilient until you've watched it fail on purpose. This talk covers the chaos experiments we ran against production, the … Show more
You don't actually know your system is resilient until you've watched it fail on purpose. This talk covers the chaos experiments we ran against production, the game-day format that got the whole team comfortable with controlled failure, and the surprising single point of failure our first experiment uncovered. You'll leave with a starter experiment you can run safely next week.
Uma Ionescu · Software Engineer, Northwind Systems
Configuration scattered across environment variables, feature flags, and three different YAML files makes debugging an environment-specific bug a scavenger hunt… Show more
Configuration scattered across environment variables, feature flags, and three different YAML files makes debugging an environment-specific bug a scavenger hunt. This talk covers how we consolidated configuration into a single typed, validated source of truth, and the migration that got legacy services onto it without downtime. You'll leave with a pattern for config validation that fails at deploy time instead of 2am.