Open to AI consulting & collaborations

I teach machines to read the fine print FDA filings financial content contracts — and to say so when they’re not sure.

I’m a Lead AI/ML Engineer with ten years in software and data, the last six building AI for industries where a wrong answer costs real money: pharma, consumer finance and law. I build RAG pipelines, agent systems and document intelligence that regulators, auditors and very busy experts can actually trust.

Regulatory AI✦Agentic RAG✦Document intelligence✦Clinical NLP✦Consumer finance AI✦Legal AI✦LLM guardrails✦Regulatory AI✦Agentic RAG✦Document intelligence✦Clinical NLP✦Consumer finance AI✦Legal AI✦LLM guardrails✦
Featured work5 selected projects
Off the clockthings I build for fun

A self-healing smart home

Home Assistant on a Mac mini with AI agents that watch, repair and explain the house — Apple Home and Alexa bridges, secure remote access through Cloudflare, and nightly disaster-recovery backups.

Home AssistantDockerClaude agentsCloudflare Tunnel

A personal AI gateway

One local gateway in front of OpenRouter and on-device Ollama models, with model aliases, a hard monthly budget, per-call logging, shared MCP tools and side-by-side model bake-offs.

LiteLLMOllamaMCPpromptfoo
AboutNoida, India

I studied statistics, started out in software and data, moved into machine learning, and for the last few years have led teams building LLM systems for regulated industries.

My favourite part of the job is the first three months on a new domain — sitting with medical writers, lawyers or product teams until I understand their work well enough to know where an AI can help, and where it absolutely shouldn’t guess.

  • 2026 —Lead AI EngineerGlobal pharmaceutical company
  • 2024 – 26Lead AI/ML EngineerAscendion — pharma, consumer finance and clinical research
  • 2024Associate Tech Lead – AI/MLKiwiTech
  • 2020 – 24Technical LeadSourceFuse
  • 2019 – 20Associate Software EngineerIGT Solutions
  • 2017 – 19Software DeveloperPremier Logic
LLMs & agents
RAG, agentic RAG, LangChain, LlamaIndex, tool calling, guardrails, evaluation, fine-tuning
ML & NLP
PyTorch, Transformers, BERT/DeBERTa, NER, XGBoost, scikit-learn
Retrieval
pgvector, FAISS, Pinecone, Qdrant, Elasticsearch, hybrid search, reranking
Platform
Python, FastAPI, Docker, Kubernetes, AWS, Azure OpenAI, RabbitMQ, Kafka, Redis
Domains
FDA/GLP, HIPAA, GDPR, consumer finance, contract law