Recent Writing
Reinforcement Learning from Human Feedback
Fri Jul 10 2026
Understanding LLM.int8(): Making Large Models Fit
Fri Jun 12 2026
Understanding Flash Attention: Making Transformers Fast
Tue Jun 09 2026
Transformer Architecture: From First Principles
Mon Jun 08 2026
Understanding Self-Attention: From First Principles
Sun Jun 07 2026
Embracing the Emptiness
Wed Dec 17 2025
Why MCP (Model Context Protocol) is Here to Stay!
Fri Apr 25 2025
Ankit Mishra
I build Software and A.I Systems from 0 -> 1, move fast, and thrive in high-ownership environments where initiative matters.
Founding Engineer at
DocuraHealth
(YC W26),
ex-MLE at
Pibit.ai
(YC W21),
and
ex-Founding Engineer at
AarogyaID
.
Curious about how it all started? Read more about me & my story, or explore my engineering work.
THE MATH THAT BENT THE TRAJECTORY OF AI
Attention(Q, K, V) = softmax(QKT/√dk)V
Lex Fridman — Go All The Way (Charles Bukowski)
0:00
Latest News
Jul 26, 2026
Update
Mutant-AI Docs Live: Behavioral Data Generation & Red Team Agents added to find vulnerabilities in AI Agents, RAGs, and LLMs (1K+ Downloads!)
Jul 22, 2026
Launched mutant-ai v0.7.3: An agentic Python package for behavioral data generation and hypothesis-driven red team attacks.