Build the retrieval, ranking, and generation systems behind AlphaLoops — the layer that turns CRM and market data into account briefs the team can rely on.
Location: London
Working closely with Tys and Tom.
Status Open
About the role
You’ll own the model layer that powers our core product: account briefs, call prep, and outbound drafts. The bar is reliability — every output is reviewed by a customer’s wholesale desk, so accuracy, traceability, and latency need to clear a high standard. You’ll work closely with Tys, the rest of engineering, and customers themselves on a system that has to perform in production, not just in a demo.
What you’ll do
- Design, build, and maintain retrieval and generation pipelines on multi‑tenant CRM and market data.
- Own the evaluation harness — accuracy, latency, cost — and the workflows that keep it honest.
- Work directly with customers to translate domain expertise into model behaviour.
- Keep outputs traceable and reviewable, with the right balance of determinism and flexibility.
What we’re looking for
- Three or more years building and shipping production ML or LLM systems.
- Experience with retrieval architectures (RAG, hybrid search, re‑rankers) on real enterprise data.
- Strong fundamentals in evaluation, observability, and data quality.
- Comfortable working closely with customers and turning their feedback into changes in the model.
Nice to have: Experience in financial services, asset management, or another regulated domain.
#J-18808-Ljbffr
