Role Overview - 3 month contract
We are seeking an experienced Data Scientist with a strong background in Generative AI to design, build, and deploy AI-powered tools end-to-end. You will work within a small, multi-disciplinary team and take full ownership of projects—from initial discovery through to production deployment. This includes scoping use cases, building prototypes, productionising solutions, and implementing robust evaluation and governance frameworks.
Key Responsibilities
* Develop and deploy Generative AI tools independently, including chat assistants, document Q&A (RAG), summarisation, classification, extraction, and agent-based workflow automation.
* Lead evaluation and safety efforts, including the creation of offline/online test sets, and measurement of faithfulness, hallucination, bias, latency, and cost. Implement guardrails and red-teaming strategies.
* Package solutions as services, APIs, or lightweight applications (e.g., Streamlit, Gradio, React), and integrate them via CI/CD pipelines.
* Design and manage data pipelines, including chunking and embedding strategies, vector store selection, prompt versioning, and monitoring for drift and quality.
* Define model strategy, selecting and combining hosted and open-source providers, fine-tuning where appropriate, and optimising for performance, cost, and privacy.
* Translate stakeholder requirements into measurable KPIs, lead discovery sessions, document solutions clearly, and ensure maintainability.
* Apply best practices in data ethics, security, and privacy, and align solutions with service standards and accessibility requirements.
Technical Environment
* Languages & Frameworks: Python (pandas, PyTorch, Transformers), SQL
* LLM Tools: LangChain, LlamaIndex (or similar)
* Vector Databases: FAISS, pgvector, Pinecone (or similar)
* Cloud & DevOps: Azure, AWS, GCP; Docker, REST APIs, GitHub Actions
* Data & MLOps: BigQuery, Snowflake, MLflow, DVC, dbt, Airflow (preferred)
* Front-End Tools: Streamlit, Gradio, basic React (for internal tools)
Required Experience
* Minimum 7 years in Data Science/ML, including hands-on delivery of Generative AI products (beyond proof-of-concept).
* Proven ability to independently deliver production-ready tools from concept to deployment.
* Strong proficiency in Python and SQL, with solid software engineering practices (testing, versioning, CI/CD).
* Practical experience with LLMs, including prompt design, retrieval-augmented generation (RAG), tool/function calling, evaluation, guardrails, and observability.
* Strong foundation in statistics and experimentation (e.g., A/B testing), with the ability to communicate impact to non-technical stakeholders.
* Experience handling sensitive data securely and in compliance with data governance and privacy standards.
Desirable Experience
* Experience working in regulated or public-sector environments.
* Familiarity with Azure OpenAI, Vertex AI, or Amazon Bedrock.
* Lightweight fine-tuning (e.g., LoRA).
* Front-end development skills for internal tooling.