Projects

Current and past research projects.

  • WorkArena & BrowserGym

    Open benchmarks and environments for evaluating web agents on real enterprise tasks. Published at ICML 2024.

    github.com

  • Apriel Model Family

    Open-source language models for enterprise AI, including Apriel-1.5, AprielReasoner, and AprielGuard.

    huggingface.co

  • BigDocs

    An open, permissively licensed dataset for training multimodal models on document and code tasks. Published at ICLR 2025.

    github.com

  • SYNTHIA

    A pioneering synthetic dataset for semantic segmentation of urban scenes, licensed to Intel, Audi, and Huawei. Published at CVPR 2016.

    synthia-dataset.net

  • AI Tools for Indigenous Languages

    NSERC-funded project building multimodal AI translation and literacy tools for the Matsigenka and Inuktitut communities.

  • Elektra Autonomous Vehicle

    An experimental autonomous vehicle platform built at UAB with full perception, planning, and control systems.