The Work
My primary focus is operationalizing Large Language Models (LLMs) to unlock unstructured data for global affiliates. I bridge the gap between state-of-the-art NLP research and production-grade commercial applications.
Architectural Focus
- RAG at Scale: Designing retrieval-augmented generation pipelines that reason over massive commercial document repositories with low latency.
- Advanced Retrieval: Implementing hybrid search (sparse/dense), semantic reranking, and context-aware chunking strategies to maximize recall.
- Document Intelligence: Converting diverse set of documents (PDFs, slides, reports) into structured, queryable knowledge bases.
Technical Stack
Core & NLP
Python
PyTorch
Transformers
vLLM
LangChain/Graph
Infrastructure & Ops
SageMaker
Bedrock
Docker
Azure
Data & Search
Vector Databases
MongoDB
ElasticSearch
SQL