Aditya Singh

I’m a B.Tech CSE (Artificial Intelligence) student in Bengaluru, focused on turning applied AI into reliable software products. I’ve shipped full-stack features and built AI-assistant experiences designed for real enterprise users.

At Eli Lilly and Company, I integrated external knowledge via Model Context Protocol with SSE streaming, and created admin governance tooling for model whitelisting and per-assistant access control. I also resolved the majority of critical/high frontend security vulnerabilities through dependency auditing and npm override strategies.

For research, I built a hierarchical multi-head EfficientNet-B0 classification pipeline on DERM12345 and established early benchmarks (98.32% binary accuracy, 0.994 AUC) on a dataset published in Scientific Data (Nature, 2024). I also ran a systematic loss-function study for 3D U-Net segmentation on BraTS 2020.

In engineering projects, I developed a Python HNSW vector search engine with benchmark harnesses, implemented a Kafka-style distributed message broker in Go with crash-recovery guarantees, and built a self-healing RAG pipeline using LLM-as-judge with graceful degradation.