Hi, I'm George. I build the data and AI plumbing that other people get to rely on. Pipelines, platforms, and lately the AI agents that sit on top of them.
The work I'm proudest of is rarely the flashy part. It's the moment a slow, expensive process turns into something simple that just runs. I've had a good few of those now. A platform that works through more than a billion LinkedIn profiles. A 70TB database move that had to arrive intact. Real-time monitoring across a 114 MWp fleet of solar farms.
At Rightmove I'm putting AI agents into the data platform so analysts can answer their own questions instead of waiting on an engineer. A few years back I left a graduate scheme at Siemens to go build things at a startup, and I've been chasing that feeling since.
Some things I've built
-
A home network you can rebuild from scratch / open sourceThe repo behind this very page, and the rest of my house. Two dozen self-hosted services, from media and photos to a local LLM and an autonomous AI agent, running on a handful of second-hand mini-PCs and Raspberry Pis, all defined in Ansible and rebuildable from a wiped disk with one reflash. GitHub Actions builds the images, a MikroTik segments everything into firewalled VLANs, and Tailscale ties it together. I open-sourced it so anyone can lift the patterns: github.com/gmourtz/gws_homelab
-
Self-serve AI agents for analysts / Rightmove, ongoingPutting AI agents into the data platform so people can get answers themselves instead of queuing for an engineer. It's part of bringing 25 years of analytical data up to date.
-
A migration that couldn't lose a row / RightmoveI built the checks that proved a 70TB move from Oracle to BigQuery was actually complete, across 25 years of history. I also got the cost of safeguarding all that data down by about a third.
-
Talent search across a billion profiles / HgI designed the pipelines behind a people-intelligence platform that works through more than a billion LinkedIn profiles, so investment teams can search for talent and find gaps across their portfolio. I handled the personal data carefully and shipped it on AWS with Terraform.
-
A faster way to model new data / Arch CapitalA metadata-driven setup on Snowflake, using Streams, Tasks and Dynamic Tables, that cut the time to model new data by about 30%.
-
Watching a fleet of solar farms in real time / GeodomiEnd-to-end monitoring for 114 MWp of solar assets, from ingestion through to the dashboards people actually used. The insights helped grow revenue by 12%. This is the startup I left the graduate scheme for.
Tools I reach for
Mostly Python and SQL, on Snowflake, BigQuery, AWS and GCP. Day to day that means dbt, Airflow, Spark, Terraform, Docker, Databricks and Meltano, plus a bit of Cypher when the data is a graph.
Certifications
- SnowPro Advanced: Data Engineer, and SnowPro Core
- Databricks Data Engineer Associate, and AWS Cloud Practitioner
- Professional Scrum Master I, dbt Fundamentals, APM PFQ
Say hello
The best way to reach me is email: [email protected].