About
I came to data science from software engineering, and it shows. A model isn't finished to me until someone else can pick it up and put it back into production.
I started as an apprentice developer in 2019 and spent five years on .NET APIs, data infrastructure and the internal frameworks other developers built on. Since 2024 I've worked independently, mostly with Havas Media, where I set up how the data science team works and built the machine learning estate behind it. Attribution, lifetime value and incrementality are the problems I keep coming back to. I keep a slip-box of what I pick up along the way.
Work
Data science & platform engineering (independent)
- Built out the data science practice: agile process, code review, MLOps, the GCP stack, templates and documentation.
- Designed and shipped a multi-client ML estate: ingestion, cleaning, user resolution, feature engineering, model application and evaluation, orchestrated across some 34 repositories.
- Production stack: BigQuery, Dataform, Dataflow (Beam), Vertex AI Pipelines, Cloud Run, MLflow, dbt on AWS Batch.
Developer
- Led development and maintenance of core company APIs and .NET data infrastructure.
- Wrote the clean-architecture template and CI/CD pipelines that the internal applications were built on.
- Mentored developers, sysadmins and technical stakeholders, and ran system administration through the pandemic alongside the day job.
Apprentice Developer
- C#, .NET, Angular, Azure, GCP, SQL and TypeScript, shipping to production from the first year.
Education
- Machine Learning Specialization
- CS50AI: Introduction to Artificial Intelligence with Python
- Higher National Certificate (Level 4), Software Development
- CS50: Introduction to Computer Science