Kiran Shrestha.

I'm an ML engineer and researcher who builds AI systems that people can actually use—from retrieval and LLM applications to computer vision, model serving, and full-stack workflow tools. My research background keeps me rigorous; my engineering focus is making the work reliable in practice.

Currently working on: Prototype-based interpretable image recognition using DINO + part-attention

Research experience

3+ years ML research

Optimized inference speedup

3.8× inference speedup

Anomaly detection result

96.3% pAUROC, 1-shot