Applied AI / Software Engineer building production RAG systems, backend AI services, and full-stack automation products.

Portfolio · LinkedIn · Email

I build software that takes AI products beyond demos into reliable, measurable systems. My work focuses on retrieval-backed applications, agent workflows, evaluation pipelines, observability, and cloud-native backend services.

Recent work includes Python and FastAPI services, TypeScript/React applications, event-driven AWS pipelines, vector search, tracing, offline evaluation, and production-oriented AI tooling.

• Retrieval-Augmented Generation (RAG) systems with grounded retrieval and measurable answer quality
• Agent workflows with tool use, routing, persisted state, and human review steps
• Backend APIs and document-processing services for AI applications
• Evaluation and observability for AI systems, including regression checks and trace-based debugging
• Full-stack delivery across Python, FastAPI, React, Next.js, PostgreSQL, and cloud infrastructure

Production-style RAG platform built with FastAPI, OpenAI, Pinecone, Ragas, Prometheus, Grafana, Docker, and AWS.

Highlights

• Built grounded retrieval and answer generation workflows
• Added offline evaluation and regression checks for retrieval and generation quality
• Improved faithfulness to 0.78 and maintained 0.75 context recall
• Built an event-driven ingestion pipeline using S3 -> SQS -> Lambda -> Fargate -> Pinecone
• Added observability for system health, tracing, and benchmarking

Repository

Multi-step legal review backend built with FastAPI, LangGraph, PostgreSQL, Pinecone, Phoenix, OpenAI, and Anthropic.

Highlights

• Built tool-connected agent workflows with human-in-the-loop review
• Improved faithfulness from 0.59 -> 0.78
• Improved answer relevancy from 0.51 -> 0.82
• Increased supported conversation length by 50% using contextual memory and session compaction
• Added persisted review runs, revision history, and trace-based debugging

Repository

Languages
Python, TypeScript, JavaScript, SQL

Frameworks
FastAPI, React, Next.js, Node.js, Express, LangChain, LangGraph

Cloud & DevOps
AWS, Docker, Terraform, GitHub Actions, CI/CD

Data & Storage
PostgreSQL, PostGIS, DynamoDB, Pinecone, Supabase

AI Tooling
OpenAI, Anthropic, RAG, evaluation workflows, observability, prompt iteration, agent systems

• Reduced repeated manual processing by 70%
• Improved application performance by 45%
• Improved planning efficiency by 15%
• Improved customer experience by 20%
• Built systems used across 5+ clients and 10+ major sites
• Supported workflows across thousands of processing events and multi-document ingestion pipelines

I have 3+ years of experience across applied AI, backend engineering, full-stack development, and Web GIS. My commercial work has covered agritech, construction, real estate, and geospatial platforms, and my recent focus has been on AI products that need reliable backend systems, evaluation, and operational visibility.

• Applied AI Engineer
• AI Software Engineer
• Backend AI Engineer
• Full-Stack AI Engineer

Based in London and open to UK opportunities.