Open to opportunities

Kirtan User

@kirtanuser

Message

I build production AI systems for enterprise retrieval, analytics, and automation.

India

Message

What I'm looking for

I'm looking to build scalable AI platforms where I can own production RAG, agentic workflows, and enterprise analytics systems while deepening my MLOps and data engineering skills in model serving, Airflow, dbt, and PySpark.

At TatvaSoft, I build enterprise AI systems spanning multi-agent RAG pipelines, persona-aware NLQ-to-SQL analytics copilots, and compliance-focused legal assistants.

I’ve shipped citation-backed retrieval workflows for legal document review on Amazon Bedrock, with knowledge-base grounding, hallucination refusal, audit trails, and structured outputs. For financially sensitive workloads, I design on-premise RAG deployments using FAISS and Ollama without external APIs.

I build Azure AI Foundry analytics copilots for 100+ ERP tables and three business personas, using FalkorDB graph-aware retrieval and sqlglot validation for safe, read-only SQL execution. By retrieving relevant schema subgraphs rather than full schemas, I reduced context size and hallucination; I also optimized production database queries by 30–40%.

I work end to end across document ingestion, embeddings, semantic retrieval, agent orchestration, FastAPI services, PostgreSQL, Docker, and Langfuse observability. I’m continuing to deepen my MLOps and data engineering capabilities in model serving, Airflow, dbt, and PySpark to build scalable AI platforms.

Experience

Work history, roles, and key accomplishments

Current

Associate Software Engineer / AI-ML Engineer

Current

TatvaSoft

Jul 2025 - Present (1 year 1 month)

Design and ship enterprise multi-agent RAG pipelines using hybrid retrieval with Cohere reranking for citation-backed knowledge retrieval, including on-premise deployments for financially sensitive workloads. Build persona-aware NLQ-to-SQL analytics copilots on Azure AI Foundry with graph-aware retrieval and query validation, and develop a Bedrock-based compliance RAG assistant with citation verif

LangGraph RAG Faiss Ollama Azure AI Foundry FalkorDB SQLGlot AWS AI Bedrock Langfuse fastAPI Docker PostgreSQL

Trainee Software Engineer

TatvaSoft

Jul 2024 - Jun 2025 (11 months)

Built and maintained backend services and REST APIs, profiling and optimizing recurring database queries to improve response times by 30-40% on flagged performance bottlenecks. Contributed to early-stage GenAI proofs-of-concept covering document ingestion, chunking, and embedding pipelines that became the basis for later production RAG systems.

Python REST APIs Database Optimization Data Ingestion Embedding Pipelines