For hiring teams
I build the plumbing
research runs on.
I spent the first part of my career analysing genomics data and the last few years building the infrastructure underneath it. Today that means CI/CD, AWS and observability for AI teams at Genentech.
Where the work
has been.
Most of my career has been in science: academic genomics research first, then regulated pharma and the data platforms behind it. I like being the person who keeps the pipelines green and makes the next team's setup easier than mine was.
What I can share here is the public version. The details live on my CV and in conversation.
AI/ML DevOps Enablement Consultant
2025 to now · Genentech, via Amaris Consulting
I run the pipelines, AWS footprint and monitoring under Genentech's internal AI projects, so the teams building them can keep shipping. That included moving an EKS cluster from Datadog to Grafana Cloud with every dashboard and alert rebuilt before the old one was switched off.
Senior Software Engineer in Test
2023 to 2024 · Roche, via Amaris Consulting
Validation frameworks for sequencing-instrument software in Java, plus a pre-commit and CI quality baseline that several repositories adopted, including security checks on infrastructure code.
Senior Scientific Data Curator
2021 to 2023 · Genentech, via Amaris Consulting
I moved more than 50 team repositories from GitHub to GitLab with their history intact, and wrote Terraform during a migration to infrastructure as code.
Biostatistician, then Bioinformatician
2018 to 2021 · Academic research
Genotyping and methylation analysis at the Douglas Mental Health University Institute in Montreal, then sequencing pipelines at the University of Calgary. I co-authored a paper in Nature Genetics (2019) along the way.
A few results
I can walk you through.
Each of these started as somebody's annoying, repeated task. I'm happy to go through how they were built, and what I'd do differently.
15 minutes down to 15 seconds
A natural-language assistant over the Ensembl genomics database. Bench scientists ask in plain English instead of hand-building API calls, which saves each of them about 6 hours a month.
An afternoon, where it used to take a sprint
A Terraform template for Grafana Cloud: four modules and a starter alert pack. A new team gets its dashboards, alerts and Slack routing from one template. It was designed for more than 35 teams.
Coding agents that can only read and comment
AI agents inside a regulated CI pipeline. One posts a daily health digest, and a second (still a proof of concept) comments on failed builds. Neither can change code. A person still makes every change.
About 5,000 people onboarded
Guides, live demos and day-to-day help for an enterprise AI workspace, plus a cross-repository promotion pattern with health checks before every deploy.
And some things
you can open today.
My side projects are where I try ideas without a ticket attached. They're older and smaller than the work above, but you can inspect every line.
Creative engineering
Training and selecting neural styles, running the batch jobs, and showing the results as images and motion.
Style-transfer case study ↗Working inside limits
Fitting a 3D cartography experience into a small hosting-memory budget, with rendered flybys and small releases.
Hinuhinu decisions ↗Interfaces for exploration
Image processing behind a mobile-friendly Shiny app.
Shiny-filteRs ↗Writing it down
Articles and talks on dimensionality reduction, APIs and reproducible workflows.
Read the technical writing ↗
Get in touch.
If something here overlaps with what your team is working on, I'd like to hear about it. Email me and I'll send a current CV.
Running a small business and curious about AI?See how I work with small teams.