Professional Experience · Research · Core Competencies
At Antz AI, I operated at the frontier of applied AI engineering — designing and deploying production-grade Retrieval-Augmented Generation (RAG) pipelines that bridge large language models with domain-specific knowledge bases. Working across the full AI stack, I integrated vector databases for high-throughput semantic search, implemented semantic chunking strategies to improve retrieval precision, and conducted systematic benchmarking to ensure low-latency, hallucination-resistant outputs ready for real-world deployment.
This research proposes a retrieval-grounded multimodal framework to address hallucinations in clinical LLMs. By unifying structured clinical text with imaging-derived contextual features, the framework ensures secure, offline, and verifiable language generation. It supports privacy-preserving deployment in resource-limited clinical settings, demonstrating high semantic consistency between retrieved evidence and generated outputs.