George Hotz started tinygrad in October 2020 and founded tiny corp, the computer company that maintains it; tinygrad is an end-to-end deep learning stack released under the MIT license. George Hotz's blog describes tiny corp's stated goal as "to commoditize the petaflop" and describes tinygrad's stack as spanning a frontend, a graph compiler, runtimes, and GPU drivers. George Hotz's homepage also points to comma.ai as one of George Hotz's startups; comma.ai develops openpilot, an MIT-licensed operating system for robotics that upgrades the driver assistance system on supported cars.123456
At a glance
- Current roles
- Founder +1 more
- Sector
- Company
- Education
- Johns Hopkins
- Work listed
- 1
- Sources cited
- 8
- Profile reviewed
Products and projects
George Hotz's homepage points to two companies Hotz started: comma.ai and tiny corp. comma.ai develops openpilot, an MIT-licensed operating system for robotics that upgrades the driver assistance system on supported cars, and sells hardware for hands-free driving in existing cars.367
In May 2023 Hotz announced tiny corp on the blog, describing it as a computer company whose goal is "to commoditize the petaflop". The company sells tinybox, a configurable deep learning computer, through tinygrad.org.18
Open-source work
Hotz started tinygrad in October 2020, with a first commit on October 17, as a small project for learning about neural networks. By 2023 it ran the driving model inside openpilot, and its intermediate representation was built from a small set of operations so that matrix multiplies and convolutions reduce to multiplies and sums with lazy evaluation allowing operation fusion.21
Hotz's December 2025 post "Five years of tinygrad" describes the stack as including a frontend, a graph compiler, runtimes and drivers, with work under way to remove the LLVM dependency so tinygrad can drive AMD GPUs in pure Python. The post describes tiny corp as working almost entirely in the open on GitHub and Discord, with contributors hired through their contributions to the repository.25
Work and contributions
Timeline
Where they work
Related people
- Ian BuckChips and hardware · ML frameworks and libraries
- Natalia GimelsheinML frameworks and libraries · Chips and hardware
- Gavin UbertiChips and hardware · ML frameworks and libraries
- Johannes GäßlerChips and hardware · ML frameworks and libraries
- Jeff JohnsonChips and hardware · ML frameworks and libraries
- Nikita ShulgaML frameworks and libraries · Chips and hardware
- Piotr BialeckiML frameworks and libraries · Chips and hardware
- Jeff BolzChips and hardware
- Harrison KinsleyML frameworks and libraries
- Chris LattnerML frameworks and libraries
Suggest a correction or removal
Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.
References
8 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026