Open to opportunities

Aryan Talati

@aryantalati

Message

I build production RAG pipelines and AI search agents with measurable retrieval quality.

United States

Message

What I'm looking for

I'm looking to build production AI search, RAG, and agent systems where I can improve retrieval quality, grounding, evaluation, latency, and cost at scale.

I've built production RAG systems and AI search agents spanning ingestion, embeddings, hybrid retrieval, graph traversal, reranking, grounding, and evaluation. At Zirozen Software Corporation, I built a Python and FAISS retrieval pipeline over approximately 5,980 records, reducing lookup time from 3–5 minutes to about 5 seconds at approximately 90% accuracy.

For SupplementsRx AI, I engineered LangGraph-based query routing, FAISS retrieval, Neo4j graph traversal, cross-encoder reranking, and RAGAS evaluation across a 250-question benchmark. I also deploy FastAPI services on AWS with Docker and CI/CD while monitoring retrieval quality, latency, drift, and inference cost.

Experience

Work history, roles, and key accomplishments

AI Engineer Intern

Zirozen Software Corporation

Jan 2024 - Jun 2024 (5 months)

Built an end-to-end RAG pipeline in Python, reducing lookup time from 3-5 minutes to ~5 seconds with ~90% accuracy. Fine-tuned DistilBERT and implemented grounding layers to prevent hallucinations, deploying REST APIs with FastAPI and monitoring.

Python RAG Faiss DistilBERT fastAPI REST APIs Monitoring