S. Keshav
I’m the Robert Sansom Professor of Computer Science at the Department of Computer Science and Technology, University of Cambridge, and a Professorial Fellow at Fitzwilliam College. I’m also a Fellow of the Royal Society, Royal Society of Canada, ACM, and IEEE.
Over the last four decades, I have designed and analysed systems of many kinds, starting with the Internet, then energy systems, and now systems for the planet. My current work is driven by the observation that climate change and biodiversity loss threaten all life on Earth, yet we have unprecedented computational and observational tools to respond. I’m convinced that computer science can—and must—be a force for planetary good.
Note: Keshav is my given name and here is how my name is pronounced: Srinivasan Keshav
Current projects
A global, open, 10-metre resolution foundation model for Earth observation. We’re making planetary-scale environmental monitoring accessible to anyone with a laptop. Tessera processes satellite imagery to detect deforestation, monitor ecosystems, and track land-use changes. It’s being adopted by conservation organisations, researchers, and governments worldwide.
Tessera users are invited to join the discussion on Zulip.
The Tessera Embeddings Explorer lets anyone browse Tessera embeddings for any 5 km × 5 km patch of land on Earth, for every year from 2017 to 2025, alongside OpenStreetMap and RGB satellite imagery. Users can label habitats, find places with similar embeddings, and evaluate classifiers on their own ground truth, with train/test splits that guard against spatial leakage. No coding needed. TEE can also generate country-level maps using user-supplied compute resources.
I’m writing a free, open-source graduate textbook. I’m using Quarto, VSCode, and modern collaborative tools to make it not just a book, but a living hub for the energy informatics community.
Browser-based labs for my Cambridge course Computer Systems Modelling, covering sampling, confidence intervals, random walks, eigenvalues, and discrete- and continuous-time Markov chains.
More details on what I’m currently working on.
Thumbnail biography
I started as a researcher in computer networking in 1988, working on topics such as congestion control, packet scheduling, disconnection-tolerant networking, and system architecture. Around 2010, after watching Al Gore’s An Inconvenient Truth, I decided to change my focus from building networks to working in the broad area of sustainability. Since then, I’ve worked on reducing carbon footprints in energy generation, transportation, and buildings, and since 2021, my focus has been on using satellite data and machine learning to understand and protect Earth’s ecosystems. I believe in serving the research community: I chaired ACM SIGCOMM from 2013 to 2017 and helped create ACM SIGENERGY and the eEnergy conference. My overall vision is to use computational thinking and computing systems to solve problems that matter.
Recent highlights
- Our paper Heatwave-induced capacity bottlenecks in European electricity grids is out in Nature Energy. With Enming Liang and Minghua Chen, we simulated hundreds of projected heatwaves for 2026–2030 on the Western European grid to find potential areas of load shedding.
- New on the blog: what’s new in TEE — global coverage from 2017–2025, a Deep MLP classifier, and new ways to check for spatial leakage in your accuracy numbers.
- New on the blog: evaluating your downstream task with Tessera embeddings — get an accuracy estimate from a ground-truth shapefile with the
tessera-evalpackage. - Just for fun: make a postcard of any point on Earth, as seen by Tessera. No science needed.
- Our Tessera paper has been published at CVPR 2026 in Denver. See the news article “Tessera arrives in Denver”.
- I’ve been elected to the Fellowship of the Royal Society.