ML / GenAI Engineer · LLM Evaluation & Benchmarking

_{Embedded on Google Kaggle Benchmarks · Bengaluru, IN}

I'm an ML/GenAI engineer focused on LLM evaluation, benchmarking, and the reasoning systems that sit on top of them. Six years in production ML/SWE, with the last 27 months embedded on the Google Kaggle Benchmarks team.

• Day job — Benchmark engineer on kaggle.com/benchmarks. Ported and validated 10+ peer-reviewed LLM benchmarks (MMLU, GPQA, FACTS, Global-MMLU-Lite, others) against published reference scores; cited contributor on the ICML 2025 Experts Benchmark; first internal consumer of the kaggle-benchmarks Python SDK.
• Building on the side — ask-my-notes, an end-to-end RAG service (FastAPI · Voyage embeddings · FAISS · Anthropic Claude). Shipped with a full cost ledger and a documented prompt-caching dead-end — receipts in the README.
• Background — MSc Artificial Intelligence, Distinction, Royal Holloway, University of London.

Shipping a new GenAI project every few weeks. Open to conversations on LLM eval, agentic systems, and applied RAG.

A working inventory — what I reach for daily, what I've shipped in past roles, and what I keep current enough to pick up when a project demands it.

Primary stack  ·  what I work in today Python · Kaggle SDK · PyTorch · Hugging Face Transformers · FastAPI · Docker · GCP (Vertex AI, BigQuery, Gemini Studio) · FAISS · Anthropic / OpenAI / Voyage / Gemini SDKs · LangChain · uv · pytest

ML & data  ·  tools I've shipped models with TensorFlow · Keras · Scikit-learn · XGBoost · Pandas · NumPy · Plotly · Seaborn · Matplotlib · OpenCV

Infra & DevOps  ·  production muscle from automation/observability roles Kubernetes · Jenkins · GitHub Actions · Linux · Bash · Grafana · Splunk · Firebase

Data stores MySQL · Oracle · BigQuery

Earlier in my career — open to picking up again C · Java · MATLAB · R · Flask · HTML / CSS · Adobe Photoshop