00 / about
Associate Software Engineer
Building AI/LLM tooling at SITA in London — codebase analysis infrastructure using Tree-sitter, Neo4j, and LLM embedding pipelines. Systems-minded, trajectory toward quantitative development.
University of Bath — MEng
Master's thesis on self-supervised ML and Physics-Informed Neural Networks for 3D transient PDEs. CUDA-accelerated.
Currently
C++ for quant dev and options pricing. Training for Venice Marathon, October 2026.
01 / experience
Where I've worked
SITA
Associate SWE · 2024 – present
AI/LLM internal tooling. Parallelisation architecture with significant inference speedup. Playwright test migration. Scrum Master for graduate team.
02 / projects
Things I've built
Dark Source Base Analyser
Codebase intelligence
Tree-sitter parsing, Neo4j graph, LLM summarisation pipeline, MCP interface for semantic codebase analysis.
C++ Options Pricer
Quantitative finance
Black-Scholes and custom neural net in C++. PINN trained on Heston PDE, Python ETL, FastAPI backend.
Homelab — aryam.dev
Self-hosted infrastructure
Hetzner VPS, Traefik, Cloudflare DNS, Tailscale overlay, Docker Compose under subdomains.
Tech News Pipeline
Automation
Azure Functions newsletter with Ollama headline classifier. Strictly breaking industry news.