Funding and Leadership

Noeri, a new artificial intelligence laboratory, has announced the closing of a $110 million funding round to develop self-improving AI systems for enterprise and scientific applications. The investment was led by Menlo Ventures, with participation from Madrona Venture Group, Kindred Ventures, Wing Venture Capital, and NEA. The round also included strategic investments from Cisco Investments, NVIDIA, and Snowflake, according to a company announcement.

The laboratory was founded by Stanford University computer science professors Yejin Choi and Carlos Guestrin. Choi, a MacArthur Fellow and former senior director at the Allen Institute for AI, serves as co-CEO alongside Guestrin. Guestrin previously co-founded Turi, which was acquired by Apple in 2016, and led Apple’s central machine learning and AI organization.

Technical Approach to Enterprise AI

Noeri’s stated mission is to move beyond “frozen” AI models—systems that remain static after their initial training phase—by developing “living systems” that self-organize and improve through real-world interaction.

The company aims to address a specific economic trade-off currently facing large organizations: the choice between high-cost, high-capability frontier models and lower-cost models with diminished performance for specialized tasks. Noeri’s technology is designed to study an enterprise’s specific data and workflows to optimize the entire system, including instructions, tools, and model parameters.

A key feature of the platform is its transparency and ownership model. The company asserts that customers retain ownership of the workflows and models they develop. Furthermore, the system is designed to provide explanations for every change made, which are then submitted to human operators for approval, similar to a code review process. The technology is compatible with both open-weight and closed-weight models.

Scientific Discovery and Research Foundations

Beyond enterprise workflows, Noeri intends to apply its technology to scientific discovery, specifically in fields such as protein and molecule research. The company suggests that smaller, specialized models can make open-ended scientific search more economically viable than using large-scale frontier models.

The laboratory’s technical framework is based on the founders’ academic research. This includes Guestrin’s work on self-distillation (SDPO) and Choi’s research into agentic orchestration and pluralistic alignment. These methods aim to allow models to learn from mistakes and coordinate complex tasks using smaller, more efficient architectures.

Expansion and Operations

Noeri is currently recruiting researchers and engineers for its offices in Palo Alto and Zurich. The company is working with a diverse group of strategic partners across various industries to ensure its technology can generalize across different sectors before specializing for individual clients.