Strange Loop
Strange Loop
Four AI researchers who just graduated from Princeton, with work spanning reinforcement learning, LLM pre- and post-training, and agentic systems shipped to enterprises and governments.
Four AI researchers who just graduated from Princeton, with work spanning reinforcement learning, LLM pre- and post-training, and agentic systems shipped to enterprises and governments.
Meet
Meet
Rhiaan, Shlok, Chirayu & Tharun
Rhiaan, Shlok, Chirayu & Tharun
Eliminating the friction between idea and experimental results for AI research.
Eliminating the friction between idea and experimental results for AI research.
The pace of research is bottlenecked not by the pace of ideas, but by the speed of implementation. Even in AI research, the design space of methods is so large that it is not possible to implement every idea imaginable.
Strange Loop is tooling that removes that friction. Reducing it lets researchers try far more ideas than they currently can: at least an order of magnitude improvement in research throughput. The end application is a web-based tool with easy ways to inspect the experiments being automated.
The team are AI researchers from labs at Princeton, so the feedback loop between customer and product is as tight as it gets. They already have a paper accepted to an ICML workshop and have automated large parts of other ML research projects in Princeton labs.
The pace of research is bottlenecked not by the pace of ideas, but by the speed of implementation. Even in AI research, the design space of methods is so large that it is not possible to implement every idea imaginable.
Strange Loop is tooling that removes that friction. Reducing it lets researchers try far more ideas than they currently can: at least an order of magnitude improvement in research throughput. The end application is a web-based tool with easy ways to inspect the experiments being automated.
The team are AI researchers from labs at Princeton, so the feedback loop between customer and product is as tight as it gets. They already have a paper accepted to an ICML workshop and have automated large parts of other ML research projects in Princeton labs.