Eiso Kant
Co-founder and co-CEO of Poolside, which builds open-weight foundation models for agentic coding
Eiso Kant is a co-founder and co-CEO of Poolside, which builds foundation models and agents for software development, including the open-weight Laguna models. Poolside traces its training approach, Reinforcement Learning from Code Execution Feedback (RLCEF), to source{d}, the company Kant built in 2016 to apply AI to code.231
Career
Research and products
Poolside's research centers on training models for software development with execution feedback: in RLCEF, models attempt tasks across many real code bases and learn from the result of running their solutions. The Model Factory, introduced in a 2025 post that lists Kant as an author, is the internal system Poolside uses to train and evaluate models.34
In April 2026 Kant and Warner announced Laguna M.1 and Laguna XS.2, the latter released as open weights on Hugging Face under Apache 2.0. Poolside's site now offers the Laguna family for agentic coding, including the on-device XS 2.1 model.21
Work
LagunaOpen weights
Co-signed the April 2026 post releasing Laguna M.1, Poolside's 225B-parameter model, and Laguna XS.2, its first open-weight model, with 33B total and 3B active parameters, under Apache 2.0.2
Model Factory (external site: poolside.ai)Proprietary product
Listed as an author of the 2025 post introducing the Model Factory, Poolside's internal framework for training, scaling and experimenting with foundation models.4
Career timeline
2016
Built source{d}, a company dedicated to applying AI to code.3
Dec 4, 2024
Co-signed Poolside's announcement of a first-party partnership with AWS.6
Jul 17, 2025
Co-authored Poolside's post introducing the Model Factory training framework.4
Nov 18, 2025
Co-signed Poolside's announcement of its acquisition of Fern Labs.5
Apr 28, 2026
Co-announced Laguna M.1 and Poolside's first open-weight model, Laguna XS.2.2
Other documented roles
Roles the sources document but that are not confirmed as current.
- Founder (2016), a company applying AI to code, source{d}3
Related people
More sourced profiles at the same organization, in the same field, or from the same school.
- Jason WarnerAlso at Poolside
- Albert JiangLanguage models
- Alessio FanelliLanguage models
- Alexander WettigLanguage models
- Antje BarthLanguage models
- Aravind SrinivasLanguage models
- Baptiste RozièreLanguage models
- Binyuan HuiLanguage models
- Bob McGrewReinforcement learning
- Craig FederighiLanguage models
- David BauDeveloper tools
- Denny BritzLanguage models
Is this you?
Profiles cover public professional work only, and every line cites a source. If something is out of date, missing or wrong, send a correction with a link to an official page and we will update it. Suggest a correction
Add this badge to your website or GitHub profile README. It links back to this page. It is not an endorsement or a rating.
Markdown (for a README)
[](https://unitedstatesofamericasuperintelligence.com/people/eiso-kant/)
HTML
<a href="https://unitedstatesofamericasuperintelligence.com/people/eiso-kant/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/profile.svg" alt="Eiso Kant — Profile on USASI" height="20"></a>
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.