Vivek Tiwari

Machine Learning Engineer (RAG Systems) at Indiana University Bloomington

Built scalable Retrieval Augmented Generation (RAG) architectures using vector stores (Pinecone) and LLMs to optimize context retrieval for 12,000+ patient queries annually. Developed data ingestion and ETL patterns to populate knowledge bases from unstructured medical text, reducing hallucination rates and ensuring high-fidelity responses for clinical decision support. Integrated GenAI scripts with AWS to serve real-time predictions, optimizing latency by 50% and adding $60K in operational value through improved system reliability.

AI Application Engineer (GenAI & Agents) at Zof AI Inc

Architected and deployed production-ready agentic systems using LangGraph and LangChain, orchestrating multi-agent QA automation workflows; reduced manual testing hours by 75% and monthly operational costs by $85K. Implemented robust LLMOps pipelines with AWS CodeBuild, ECR, and Docker for containerization and CI/CD deployment of GenAI services with 99.5% system uptime. Designed intelligent tool-calling mechanisms within AI agents, integrating human-in-the-loop feedback to improve decision accuracy by 30% and accelerate client adoption of generative AI solutions.

Data Scientist at Indiana University Bloomington

Engineered distributed data pipelines using Spark to clean and transform 1.2M+ documents, creating a structured knowledge base for downstream NLP and classification tasks. Collaborated with cross-functional teams to operationalize ML models, designing interactive dashboards that translated complex model evaluations into actionable business insights. Implemented automated data quality checks within the ingestion pipeline, ensuring 99% data integrity and streamlining the training process for large-scale document analysis models.