ML Engineer with 2.6 years taking AI from idea to production — RAG pipelines, LLM applications, NLP systems, and MLOps infrastructure. I write the code, deploy the model, monitor the drift, and keep it running.
→ Currently focused on Production ML for Fintech & Banking.
I'm a Software Engineer specialising in ML/AI — I've spent 2.6 years building systems that actually run in production, not just experiments that live in notebooks. My stack spans RAG pipelines, LangChain, LLM fine-tuning, NLP, and MLOps on AWS and Azure. I'm now deliberately targeting ML Engineer roles in Fintech and Banking — actively studying AML systems, fraud detection patterns, and financial ML to complement my production engineering background. If you need someone who ships reliable AI and learns domain fast, let's talk.
2.6 years of building and deploying real AI systems — RAG, NLP, computer vision, MLOps — across multiple production environments
Consistent upskilling — MCP agents, GenAI strategy, and foundational AI/ML
Completed: January 10, 2026 · 8 hours
Completed: April 26, 2021 · 16.5 hours
Completed: January 21, 2026
Completed: January 22, 2026
Full ML engineering stack — model development, deployment, monitoring, and infrastructure
AI systems I built, shipped, and maintain — not toy demos
RAG pipeline that lets users query car insurance PDFs in plain English. Chunks policy documents semantically, stores in vector DB, retrieves relevant clauses and generates grounded answers via LLM — no hallucination beyond source material. Built with LangChain, deployed on custom domain.
Institutional-grade trading system that computes live Greeks (Gamma, Delta, Theta) from NSE options chain.
Enterprise RAG chatbot on LangChain + Azure Cosmos DB — covers 22+ customer topics with sub-3-second latency. Uses Few-Shot prompting and Chain-of-Thought reasoning for structured, reliable answers at production scale.
Multi-head DistilBERT model in production — classifies 22+ complaint categories at 95% real-time accuracy on live multi-source data. Anomaly detection layer reduced manual monitoring burden by 80%.
XGBoost + Random Forest ensemble predicting repeat callers from behavioural patterns. Deployed to production at a contact centre — delivered 60% cost reduction through smarter resource allocation.
YOLO-based real-time pipeline detecting license plates and helmet compliance — shipped inside the Emotyx product. End-to-end CV system: ingestion, inference, alerting, and reporting in production.
NLP pipeline that extracts and clusters 200+ pain points from raw customer feedback using semantic similarity and transformer embeddings. Replaced weeks of manual analysis with an automated daily pipeline.
Open to ML Engineer and AI Engineer roles — Fintech, Banking, and high-growth startups preferred