My story
I started breaking things in class 9, rooting a Samsung phone and flashing custom Android ROMs to see what would happen. That curiosity carried me through an Integrated M.Sc. in Mathematics at NIT Patna (CGPA 8.89/10), where my master's thesis on ML and blockchain for healthcare insurance fraud detection won the Best Final Year Project Award; my first IEEE paper (IC3SE 2024) was on GAN style transfer.
I then worked as an engineer at Eut Hub, Arvo, two stealth AI startups and Insight Labs AI, building agent systems, data pipelines and full-stack AI products. In April 2025 I founded VibeTensor, a DPIIT-recognised startup, and built Attestix, an open-source identity and attestation layer for AI agents (12,000+ PyPI downloads all time per pepy.tech, September 2026).
Around the same time I started doing AI safety work properly: first as a participant in METR's randomised trial on AI coding tools, then through four BlueDot Impact cohorts. In 2026 I moved to London for the Iliad alignment-theory intensive at LISA, and then LASR Labs, where I am a researcher in the Summer 2026 cohort, supervised by Satvik Golechha (UK AI Security Institute). Our team works on the science of model organisms. Most of what I care about is measurement: whether a reported number is a property of the model or of the scoring pipeline.