Aadish Sanghvi

I’m a final-year Computer Science undergraduate specializing in Data Science, with hands-on experience building ML models and agentic AI frameworks. I enjoy turning messy, real-world data into reliable systems that can be monitored, audited, and improved.

As a Data Science Intern at Bajaj Finserv Health, I’ve built production-grade pipelines for intent extraction, schema grounding, and SQL generation. I also developed an auditing approach to track daily ETL failures with error logs and reusable artifacts for GENAI/NLP workflows.

My work blends modern NLP and MLOps: I’ve implemented RAG-style components, fine-tuning with PEFT/LoRA/QLoRA, and orchestration using LangGraph and LangChain. For observability and performance evaluation, I’ve set up Langfuse tracing and used MLflow, while leveraging vector stores like FAISS, Qdrant, and Pinecone.

I’m driven by shipping measurable outcomes—whether that’s building multi-agent solutions like MediMind (multi-agent medical diagnosis) or creating fraud detection and optimization systems with strong accuracy. I’m especially interested in teams where I can keep expanding from model-building to end-to-end AI engineering: data pipelines, orchestration, and deployment with strong engineering discipline.