Principal AI Engineer at Blue Yonder with 13+ years as a software engineer, building scaled, reliable systems. I started as a full-stack developer, and these days I spend most of my time planning AI solutions and getting them into production. I'm based in the EU and can work on site in both the UK and the EU.
I work on the infrastructure that makes AI systems dependable once real users depend on them:
- Agentic systems. I design and build custom agentic frameworks, with the orchestration, tool use and failure handling that production agents need.
- Observability. I run LLM tracing and prompt management on Langfuse, so every agent run can be inspected after the fact.
- Analytics. I use ClickHouse to query trace and usage data at volume.
- Evals. I build evaluation suites and use them to decide whether a change ships.
On the full-stack side I worked mainly in Django, with Vue and Nuxt on the front end. Thirteen years of shipping systems that had to stay up under load still shape how I work. I can read a schema, trace a request through a backend and talk to the people building the front end, which helps when an AI feature has to fit into an existing product.
Most of my current work is in private repositories due to data privacy.
- Site: aasimkhan.dev
- LinkedIn: aasim-mk
- X: @aasim_mk



