Patrick Beukema

Hi! I’m Patrick, a scientist and engineer passionate about building AI for social good at scale. I lead a new AI team at Ai2 focused on large-scale, mixed-modality planetary intelligence for humanitarian and environmental applications. We work across many areas including sustainability, climate change, wildfire forecasting, agriculture, food security, and beyond. To make this intelligence as useful as possible, we are building a new open platform called earth-system. We are hiring research scientists and engineers–if you’re excited to advance the state of the art in this space, please reach out! See: https://allenai.org/careers1.
Beyond my work at Ai2, I advise non-profits on how to leverage AI for science, sustainability, and biomedical research, and I also consult for hedge funds on the future of AI. I host a monthly seminar at Ai2 on applied AI, many of these talks are available on YouTube. Outside of AI, I’ve had the opportunity to help found a medical clinic in Haiti and a bicycle co-operative in Montréal.
Before joining Ai2, I led a variety of zero-to-one applied ML products across startups, tech, and academia (CMU, McGill, and the University of Pittsburgh). During my Ph.D. I focused on brain decoding and learning at the CNUP and CNBC. Afterwards, I joined a startup and built real-time neural decoding models and streaming computer vision systems for facial recognition and object detection (acquisition press release). During my master’s at CMU, I explored recurrent neural networks and causal inference. Before graduate school, I studied math and logic at McGill and spent a lot of time riding a Marinoni through Montréal’s icy streets. I have also worked on hybrid CV-NLP modeling. I have published a little across neuroscience, biomedical research, and applied machine learning.
I now live in the majestic Pacific Northwest with my wife and our two young boys. Lately I’ve been exploring the science of home espresso, which is proving quite useful :).
-
Note to prospective applicants: we receive an overwhelming number of applications for open roles. Regretfully, I cannot respond to every inquiry. ↩
Jun 25, 2025 | New paper on modeling wildfire risk (@ICML-TerraBytes proceedings) [GitHub] |
Jun 11, 2025 | CVPR Keynote: Closing the Gap: Building Planetary AI for Real-World Users |
Jun 1, 2025 | “Satellite imagery and AI” V2.0 now covering Landsat, VIIRS, S1, and S2! |
May 27, 2025 | Hiring Research Engineers to build Earth-System. |
May 15, 2025 | “Galileo” a new multimodal EO FM accepted at ICML! [blog][paper][code] |
Mar 6, 2025 | “Atlantes” spotlight at ICLR CCAI! [GitHub][arXiv][blog][YouTube] |
Mar 1, 2025 | “Lighthouse” accepted at ICLR ML4RS! GitHub, arXiv |
Feb 13, 2025 | New paper on foundation modeling, research led by Gabriel Tseng, arXiv |
Feb 5, 2025 | Yawen Zhang from our lab to present at the 1st NASA/ESA Foundation Modeling meeting |
Feb 1, 2025 | Hiring multiple roles (Scientists/Engineers) to build and scale earth-system! |
Dec 1, 2024 | Hiring an AI intern to build next gen multi-modal foundation models for earth! |
May 30, 2024 | “Beyond AGI: Building AI for the Planet” Thursday, May 30, 2024, 4 – 6 p.m. |
Apr 18, 2024 | Interview on the MLOps podcast: Can AI help save the planet |
Mar 15, 2024 | Blog post: Accelerating Engineering with LLMS |
Feb 25, 2024 | We have open positions for an AI engineer and an AI intern! |
Jan 26, 2024 | Invited talk on “Data and High Performance AI” at ESIP |
Jan 1, 2024 | Recognized as one of the top 100 people in AI by Business Insider. |
Dec 17, 2023 | Best paper award at NeurIPS Computational Sustainability |