I design, build, and ship production-grade artificial intelligence systems with robust evaluation, observability, and reliability. My work focuses on LLM-driven products (RAG, NLQ), scalable APIs, and repeatable evaluation infrastructure grounded in real-world constraints (latency, cost, SLOs, guardrails).
- End-to-end LLM systems (RAG, NLP query pipelines)
- Evaluation harnesses & metrics dashboards
- Production observability, monitoring, guardrails
- Model serving, APIs, structured logging
Enterprise-grade ML Systems
-
End-to-end NLQ → entity resolution → GraphQL execution pipeline
- Contract-first OpenAPI 3.1, RBAC, production deployment
- p50/p95/p99 latency metrics and service level objectives
-
Evaluation harness for agentic RAG systems
- 2,600+ eval runs, sliceable dashboards, CI schema gates
-
FastAPI LLM match insights service
- Multi-provider routing, rate limits, cost guardrails
MSc in AI at the University of Edinburgh,
with experience in LLM tooling, 3D generative models, and real-world NLP/CV projects.
Actively seeking roles in:
- ML Engineering · Applied AI Research · Generative AI
Additional Projects
Lightweight demos, class presentations, and collaborations:
• simulation_enviroments_guest_lecture – Guest lecture repo showing RL environments (DQN & SAC)
• MAGNN_RecipeRecs – Collaborator on MAGNN-based recipe recommender
• CLIP_Segmentation_Demo – Collaborator on zero-shot segmentation using CLIP