sqln User

I build production-ready AI systems that turn complex data into reliable decisions. As an AI Engineer, I developed machine learning pipelines for GPU fleet health monitoring by analyzing large-scale telemetry streams and identifying anomalous GPU behavior.

I also designed predictive models for GPU failure detection and real-time health scoring, supporting proactive infrastructure maintenance. I developed FastAPI inference services for real-time anomaly detection and model serving, collaborating with engineering teams to deploy scalable AI services for production datacenter environments.

My work centers on LLM orchestration and evidence-grounded generation. I’ve engineered scalable document ingestion and vector indexing pipelines for low-latency semantic search, and improved retrieval quality using hybrid retrieval (FAISS + BM25) and Cross-Encoder reranking.

I’m especially motivated by agentic workflows and evaluation-driven development—building modular LangGraph agents with human-in-the-loop review and deploying via MLflow, Docker, and FastAPI. From agentic clinical decision support (including uncertainty estimation and Grad-CAM explainability) to medical RAG and an AI evaluation/benchmarking platform, I focus on shipping systems that are both effective and verifiable.