Skip to main content

Explore the research behind trustworthy AI

What We Work On

  • Jet engine turbine spinning, in motion blur

    01Formal Methods

    Mathematical proof, grounded in real code.

  • Industrial robot arm sweeping across the frame, in motion blur

    02AI

    Making AI useful for hard engineering problems

  • A corridor of server racks in a data centre, lights blurred

    03Accelerators

    Performance without sacrificing correctness

How We Work

  1. 01

    Apply research to real systems

    Our ideas are grounded in real engineering problems, real codebases, and systems where correctness, performance, and trust matter. We move beyond theory by applying research where the consequences of software failure are real.

  2. 02

    Publish work that moves the field

    We contribute to the research community by sharing findings, benchmarks, technical reports, and open-source tools that advance the state of the field. Our researchers publish, present, and collaborate across papers, conferences, seminars, and academic partnerships.

  3. 03

    Learn and lead from the frontier

    We bring together researchers, engineers, and domain experts across AI, programming languages, systems, and formal methods to challenge assumptions and advance what high-assurance software can become.

Research

featured

Back in the Race: US Open-Weight Models Closing the Performance Gap

On an internal benchmark translating C++ to Rust, Muse Spark Open reached the top tier of judged translation quality, and strong results across four US-developed model families point to more promising options for teams that need US-developed models to run on their own infrastructure.

read on
Two scatter plots of functions judged correct against normalized time per task, one for US-developed open-weight models and one for Chinese-developed reference models, showing US-developed models reaching the top of the benchmark.

Formally Speaking

  • Ben Greenman

    The University of UtahTyping Python, Gradually

  • John Regehr

    UtahTranslation Validation for LLVM's AArch64 and RISC-V Backends

  • Adam Chlipala

    MITScaling Formal Verification to Complete Hardware-Software Stacks

Advance the frontier with us

We're hiring researchers, engineers, and builders to bring trustworthy AI to systems that can't fail

open roles