Karthik Mukkavilli


Karthik Mukkavilli


AI for Climate Research Scientist


IBM Research Europe - Zurich Zurich, Switzerland


In the first 18 months at IBM, Dr. S. Karthik Mukkavilli worked on projects recognized with IBM's Outstanding Research Accomplishment for "Pioneering Foundation Models and Generative AI for Accelerated Discovery", [1, 2, 3, 4] involving open source models and open science in climate and sustainability. He also contributed to "Geospatial Discovery Network" recognised with IBM's A-level Research Accomplishment. He worked on research contracts that generated revenue (from space, government, energy, and environmental agencies). He works on foundation models, generative AI, geospatial, weather and climate projects with collaborators like NASA, STFC UK, ExxonMobil, Oak Ridge National Lab, and meteorological agencies like ECCC.

During his postdoc at Mila with Turing Laureate Professor Yoshua Bengio, he helped realign the AI community towards climate and Earth sciences, including work on generative AI for visualizing climate change [1, 2]. He founded and chaired workshops at ICLR, and NeurIPS on AI for Earth Sciences, forming a network of over 250 interdisciplinary researchers, supported by advisors such as Greg Dudek, Aaron Courville amongst others. He helped co-author the paper Tackling Climate Change with Machine Learning cited 1000+ times, with his Postdoc advisor Bengio, and AI leaders like Andrew Ng, and Demmis Hassabis, and was an invited speaker and organiser at the first workshop on Climate Change AI launched at ICML in 2019. Prior to Mila, Dr. Mukkavilli worked at McGill's Mobile Robotics Lab, where he pioneered early initiatives from Montreal to bridge gaps between AI and Environmental sciences, chairing sessions like EnviroNet at American Meteorological Society AI conferences [1, 2, 3, 4].

His PhD was at UNSW in the School of Photovoltaic and Renewable Energy (founded by FRS Prof. Martin Green), focusing on solar forecasting and aerosol optics for the Australian energy grid and market operator. During his PhD, he was a Commonwealth Office of Chief Executive Scholar at CSIRO, a visiting PhD affiliate at the Harvard-Smithsonian Center and contributed to "Drawdown", a New York Times Bestseller, as a research co-author focusing on plant-rich diets and carbon cycle modeling.

Dr. Mukkavilli's co-authored works have been cited over 2100+ times and highlighted by NASA [1, 2, 3], MIT Tech Review [1, 2], National Geographic, Times of India, Indian Express, CBC, New York Times, HPC Wire; received coverage in Bloomberg, Forbes and Engadget; and Dr Mukkavilli has been interviewed by Le Devoir. He has reviewed 500+ climate projects and companies, for Keeling Curve Prize. He holds a Masters from Imperial College London in Chemical Engineering and Process Systems Engineering and held an exchange scholarship at UCLA, during his undergrad in electrical and computer engineering at The University of Auckland.

Dr. Mukkavilli was co-founder, CSO at vayuh.ai, a climate-tech AI startup out of Berkeley, providing forecasts for a major US bank's commodity trading desk, after working as a project scientist, on exascale computing projects with DOE Labs (LBNL, PNNL) and the University of California [1, 2]. He co-founded and led groundobs, a creative destruction lab finalist after his PhD.


Top collaborators

Peter Staar

Peter Staar

Principal RSM; Master Inventor; Manager of `AI for Knowledge` group.