Rhythm Garg
Rhythm Garg is an artificial intelligence researcher and entrepreneur who worked on reasoning models at OpenAI and was listed as a core contributor to the o1 model, and who later co-founded Applied Compute, an AI company that helps businesses train and improve models they own, with Yash Patil and Linden LiLinden LiLinden Li is a machine learning systems engineer and entrepreneur who worked at OpenAI on machine learning systems and infrastructure for reinforcement learning training, and later co-founded Applied Compute, an AI company that trains and deploys specialized models for enterprises, with Yash Patil and Rhythm Garg.Open the full entry.
Contents9 sections
Key facts
2024Contributing to OpenAI’s o1 Reasoning Model
2025Leaving OpenAI to Co-Found Applied Compute
According to ForbesForbesForbes is an American business media brand that began with Forbes magazine, founded by B.C. Forbes in 1917. It is known for rankings such as the Forbes 400, the World’s Billionaires list and 30 Under 30, and operates Forbes.com, ForbesLive events and licensed editions in 81 countries. Since 2014 a majority stake has been held by the Hong Kong investment group Integrated Whale Media; a 2021 plan to list on the New York Stock Exchange through a special purpose acquisition company was terminated in 2022. It is headquartered in Jersey City, New Jersey.Open the full entry, he met Yash Patil and Linden Li at Stanford; after Patil joined OpenAI, he referred Garg, who worked on reasoning models there and in turn referred Li. The three then left OpenAI to found Applied Compute, which closed a $20 million seed round in June 2025.[5]
October 29, 2025Applied Compute’s Public Debut
On October 29, 2025, he and his two co-founders signed a post introducing Applied Compute, which said the company had $80 million in funding from investors including Benchmark, Sequoia, Lux and Elad Gil. The company’s website describes its business as helping companies train, serve and continuously improve models they own.[3][6]
April 8, 2026Closing a New $80 Million Round
On April 8, 2026, a company post signed by him and the other two co-founders announced that Applied Compute had raised $80 million led by Kleiner Perkins at a $1.3 billion post-money valuation, bringing its total funding to $160 million.[7]
June 2026Publishing Research Posts on Agentic Routing and Reinforcement Learning
On June 4, 2026, he and Nic Becker published a company blog post describing how they trained a small open-source model to choose which of three frontier coding models should handle each software engineering task, weighing success rate against cost. On June 16 of that month, he co-wrote another post with Nic Becker and Jacob Philips on continued training with entropy-preserving reinforcement learning.[8][9]
September 2026Applied Compute Reported in Talks for New Funding
Forbes reported in September 2026 that Applied Compute was in talks to raise $350 million at a $3.25 billion valuation, led by Elad Gil; the company cautioned that the figures could change before the round was finalized.[5]
October 9, 2026Introducing Decision Model Training in AC2
On October 9, 2026, he published a post introducing decision model training in the company’s AC2 platform, which adds a small output layer to an open-source language model so that it assigns a probability to each option in a fixed set.[10]
Related Organizations, Websites, and Public Accounts
OpenAI: Official website: https://openai.com/ (opens in a new window)[2]
Applied Compute: Official website: https://www.appliedcompute.com/ (opens in a new window)[6]
Sources
- Learning to reason with LLMs (opens in a new window)
- OpenAI o1 Contributions (opens in a new window)
- It’s time to get specific (opens in a new window)
- OpenAI o1 System Card (opens in a new window)
- Meet the ex-OpenAI guys building a cheaper open source alternative (opens in a new window)
- Applied Compute (opens in a new window)
- The Advantage You Own (opens in a new window)
- Training an Agentic Router for Optimal Cost-Performance on SWE Tasks (opens in a new window)
- Blog (opens in a new window)
- Train your own decision model in AC2 (opens in a new window)