and putting LLMs in the hands of engineers and business teams across a FTSE 250 company. Obsessed with what comes next: AI that learns from the people using it, not just the data it was trained on.
Building and maintaining GenAI infrastructure. Enabling development teams with tools and frameworks to utilise AI capabilities across the business.
FastAPILiteLLMLlamaIndexAgentCoreAWSDocker
2024 2025
Software Engineer (Data Science)
Health & Safety Executive · Liverpool, UK
Built and deployed production NLP models and RAG pipelines on large-scale risk and dangerous occurrence data, with outputs directly informing health and safety policy decisions.
LangchainPyTorchCUDAFAISSNLP
2022 2024
Analyst
Discovery Education · London, UK
Built data pipelines and semantic search infrastructure over 2.5M+ user records, enabling personalised content recommendations and downstream analytics.
PythonpandasSQLElasticsearchPrefect
Selected Projects
3 featured
01
RAG Evaluator
2025 Open Source
A framework for benchmarking retrieval strategies across document corpora. Supports FAISS, Pinecone, and Weaviate with configurable scoring metrics.
94%
recall@5
3×
faster eval
PythonFAISSLlamaIndexPytest
02
LLM Guard
2024 Internal Tool
Prompt injection and PII detection middleware for production LLM APIs. Deployed as a FastAPI service with configurable policy rules.
PythonFastAPITransformersRedis
03
Context Compressor
2024 Research
Investigated model-based context compression for long-context LLMs. Achieved 60% token reduction with less than 3% quality degradation on QA benchmarks.