Heyoo, I'm Finn.
I find out how things break
before you do.
Infrastructure, incidents, and everything in between.
PANIK
An internal tool for running incidents from Slack. Most of the architecture behind it: API Gateway, Lambda, SQS FIFO queues, deliberately no database, so there's nothing to maintain at 3am. A number of the features on top, including the pinned incident message that tracks roles and stage and stays in step with PagerDuty. Plenty of hurdles along the way, from Slack's rate limits to queue semantics to a rocky first release.
cdk-base
Every service that deployed with CDK had its own copy of the dependency tree, so when a vulnerability landed in the ecosystem, every team had to patch it separately, and the busy ones didn't. cdk-base is one hardened set of those dependencies, published as an npm package, a Docker image and a Devbox plugin. Patch it once and every consumer picks it up. Three service templates now build from it, and the image rebuilds daily.
Murmur, CloudFormation to CDK
Murmur is Culture Amp's largest service, and parts of its infrastructure were still on hand-written CloudFormation from years back. Moved onto CDK with no downtime: feature flags to stage each step, existing resources imported rather than recreated, a migration guide so the next person could repeat it. The database layer across every environment, then the main web stack.
Decommissioning staging and dev farms
Retiring the staging environments and a long list of development farms that nobody used but everybody paid for. Around a dozen repos, the shared infrastructure config, networking, databases, and the pipelines that deployed to them. Saved the company a meaningful amount of money.
Product Recovery
The annual exercise of proving we could rebuild the product after losing a region. It costs engineers a lot of time each year, and I'm working on improving it: running interviews with teams across engineering, collecting evidence on where the time actually goes, and taking that into the architecture discussions that are reshaping how we run it.
Journey
Before Culture Amp.
Aug 2023 – Feb 2024 · internship
Compass Education
Software Engineer
Compass is a school management system that 100+ schools use every day. Worked across it: the React Native mobile app, the C# and .NET services behind it, and the Ext.js web front end. Sped up the API calls the app makes on load, moved legacy data access onto the newer pattern, and built the Favourite News feature. About 20% faster loads, about 30% quicker retrieval.
- React Native
- C#
- .NET
- ASP.NET Core
- MongoDB
- Ext.js
Jul 2024 – Dec 2024 · internship
TinyTix
Full Stack Developer
An event platform, end to end. The event site, the Postgres schema behind it, which took 5,000-odd submissions, and a Dialogflow search for finding events, all deployed on AWS. Engagement went up about 40% once the new site was live.
- React
- Node.js
- PostgreSQL
- AWS
- Dialogflow
Feb 2025 – now · full-time
Culture Amp
Associate Site Reliability Engineer
All the pipes above, and on call for anything around them.
- AWS CDK
- TypeScript
- Docker
- Buildkite
- Backstage
- PagerDuty
- MongoDB Atlas
- Datadog
About
Outside of work.
Runs a lot, plays badminton, keeps a small dog called Odie and a five-minute sand timer on the desk. Snacks between tasks and works better for it. On to a marathon soon. When something doesn't make sense: reread, break it into pieces, find whoever holds the most context. Still touches grass, to general surprise.