We are looking for a Senior Data Scientist to join a seed-stage AI project on a full-time contract basis (2–3 months, with possible extension or transition to a permanent role). The project is building a computable model of the global chemicals and materials economy.
This is a hands-on role — you will be setting the standard, not following one.
Location: London, hybrid.
Start: ASAP.
Format: Contract, full-time, 2–3 months.
Your responsibilities will include:
- Profile and explore complex real-world industrial datasets — chemical plants, supply chains, material grades.
- Build baselines, then define evaluation sets and metrics before modelling.
- Ship classifiers and matching/inference models with measured error rates and written error analyses.
- Use LLMs where they genuinely belong — extraction, constrained labelling, code generation — with proper evals, never open-loop guessing.
- Document model decisions, failure modes, and validation approaches clearly.
What we expect from you:
- Proven track record of shipping models to production where being wrong had real consequences — and ability to explain exactly how they were validated.
- Strong Python — pandas, scikit-learn, solid statistical instincts.
- Ability to define evaluation frameworks and metrics independently, not just implement someone else's.
- An independent operator mindset — comfortable setting standards in an early-stage environment.
- Experience building LLM pipelines with real evaluation harnesses is a strong plus.
- No chemistry background needed — domain expertise is provided by the team.
Soft skills / Mindset:
- Rigorous — defines what "good enough" means before building, not after.
- Communicates clearly — can explain model decisions, tradeoffs, and failure modes to a non-technical audience.
- Fast — comfortable with startup pace and ambiguity.
- Honest about uncertainty — flags unknowns early rather than papering over them.
Pay: £3,000.00-£5,000.00 per month
Work Location: Hybrid remote in London