Data Scientist

Role

Coefficient is hiring a full-time Data Scientist to design, build, and deploy AI and machine-learning solutions for a range of clients (including UK government and startups). The role combines hands-on model development, productionisation, and delivery in a consultancy setting, alongside contributing to internal product work.

Key Responsibilities

• Deliver end-to-end data science and engineering solutions for client projects and internal products.
• Build, evaluate, and productionise ML and LLM systems (versioning, monitoring, iterative improvements).
• Implement LLM application patterns such as Retrieval-Augmented Generation (RAG), fine-tuning, and LLM evaluation suites with a focus on measuring quality (hallucination/faithfulness).
• Work with embeddings and vector databases, prompt engineering, and frontier model APIs (OpenAI, Anthropic, Google).
• Use AI-assisted development tools (Claude Code, Cursor, GitHub Copilot, Windsurf, Codex) in spec-driven or multi-agent workflows; critically review and correct generated outputs.
• Share knowledge through mentoring, presentations, and client-facing materials (slide decks, sprint demos).

Requirements

• Fluency in Python and SQL for data analysis, machine learning, and experimentation.
• Experience with the Python data stack: pandas, NumPy, scikit-learn, matplotlib, Seaborn.
• Hands-on experience with at least one LLM application pattern (RAG, fine-tuning, or building evaluation suites) and an emphasis on measuring model quality.
• Familiarity with embeddings, vector databases, prompt engineering, and model APIs (OpenAI, Anthropic, Google).
• Experience thinking about production concerns (MLOps / LLMOps): versioning, monitoring, and iteration beyond notebooks.
• Strong communication skills and willingness to mentor and present work to clients and internal stakeholders.
• Right to work in the UK is required; visa sponsorship is not available. Candidates in full-time education are not eligible.

Location & Eligibility

• UK-based, hybrid role. Team is based in Central London; remote-friendly but requires ability to travel to London approximately 2–4 days per month.
• May be required to work on-site at clients’ offices.

Compensation & Benefits

• Salary: £38,000 per year with a performance-based uplift after successful 3-month probation.
• 33 days annual paid holiday (including bank holidays).
• Pension contributions via Smart Pension.
• Conference budget: £1,000 in year 1, rising to £2,000 by year 3.
• Mental health support (Spill) with 8 hours of therapy covered annually (up to £520) and Headspace membership.
• 10% Time (4 hours per week) for skill development or side projects.
• Company-provided Apple laptop and peripherals.
• Access to co-working days and Hubble co-working network; paid learning and professional development.

Hiring Process

• Round 1: Application screening (CV, screening questions, code samples required).
• Round 2: One-way video (non-technical).
• Round 3: Practical coding exercise (1 hour).
• Round 4: Technical interview with pair programming (1 hour).
• Round 5: Final conversation with the CEO.