At Capgemini, I build data and AI solutions on a foundation of ETL and cloud data warehouse delivery. As an Analyst, I built and maintained client data pipelines with Informatica PowerCenter and IICS, loaded data into Snowflake, and automated routine jobs with shell scripts.
Outside work, I build and publish Python data and AI systems with tests and CI. My projects include ReData, an incremental ingestion engine on DuckDB; ETLGuard, a validation and load toolkit with SCD Type 2 history and 91% test coverage; and a tenant-isolated RAG platform with hybrid retrieval and citation checks.
Most recently I built Catalog Graph RAG (github.com/theraghavraman/graph-rag-model): an LLM planner turns data-lineage questions into validated queries that a DuckDB graph answers exactly. Benchmarked on 1,080 scored answers with local Llama 3.2 3B and Qwen2.5 Coder models, it gave 53-72% exact answers against 19-35% for BM25 and vector RAG baselines. I am looking for Data Engineer and AI Engineer roles.

