Skip to main content
P

Pathway

Pathway develops the BDH post-transformer AI model addressing continual learning, paired with a streaming data processing engine for transparent, contextualized enterprise intelligence.

Pathway, founded in 2020 and headquartered in Palo Alto, California, develops a post-transformer frontier AI model alongside a streaming data processing engine. Its BDH architecture is described as the world's first post-transformer frontier model to solve continual learning, and the company states that it outperforms traditional Transformer models. The accompanying data processing engine is billed as the fastest on the market; the two together are intended to deliver contextualized, experience-driven intelligence to enterprises, with an emphasis on transparency and full visibility into how the model works.

The company's technical scope spans continual learning, post-transformer neural architectures, attention mechanisms, speech recognition, complexity science and quantum computing, applied to enterprise and explainable AI. Its client base includes NATO, La Poste and Formula 1 racing teams, spanning defence, logistics and motorsport. Beyond Palo Alto, Pathway maintains a presence in Paris and Wroclaw.

The team was founded by Zuzanna Stamirowska, a complexity scientist. CTO Jan Chorowski worked with Nobel laureate Geoff Hinton at Google Brain and was the first to apply attention to speech; CSO Adrian Kosowski is a computer scientist and quantum physicist who completed his PhD at the age of 20. Investors include TQ Ventures and Lukasz Kaiser, co-author of the Transformer architecture.

Open jobs

No open jobs right now.