I try to mitigate climate change using computer science. I co-founded Open Climate Fix, a non-profit research lab focused on reducing greenhouse gas emissions. Previously, I was a Research Engineer at DeepMind, where I used machine learning to predict wind power.
Since I was a kid, I’ve been interested in how biological systems learn & process information; and I’ve been keen to replicate this in machines. I studied Neuroscience BSc at UCL & absolutely loved it.
But, towards the end of my BSc, I became increasingly worried about climate change (I’m now terrified by climate change). I couldn’t see how to solve climate change using neuroscience, so I said goodbye to neuroscience.
I had been working on films since my teens (as a hobby) so it seemed natural to try to make a living making environmental films. I had a lot of fun working on a wide variety of films; but it turned out to be almost impossible to make a living making environmental films.
In 2010, I went back to my geeky AI-inspired roots and did the Computing Science MSc at Imperial, with the explicit aim of using AI to mitigate climate change. The MSc was a huge amount of fun! For my end-of-year research project, I proposed an energy disaggregation project, which I really enjoyed so I applied to do a PhD on energy disaggregation. After my PhD, I did a post-doc (based at Imperial but paid for by the lovely folks at EDF Energy) looking at how to help the disaggregation field evolve rapidly.
During my PhD, I also became fascinated by getting kids excited about engineering (by running “making” workshops). I also fell in love with open-science.
However, the last paper I wrote during my PhD showed that there really isn’t much good evidence that disaggregated energy bills encourage people to save energy. This was very disappointing because, by this point in my life, I was convinced that I had to focus my career on reducing greenhouse gas emissions as fast as possible. So I decided to shift my focus away from disaggregation.
I did lots of research into how best to mitigate climate change using computer science. At points I was quite down-beat: hardly any of the options I considered had much chance of having a sizeable real-world impact.
In late-2016, as part of my research into what to work on next, I sent an email to a researcher at DeepMind asking for his opinions on how best to mitigate climate change using computer science. I was extremely fortunate that this email exchange led to me entering the DeepMind recruitment process with the result that I landed a job at DeepMind, working on energy problems. I thoroughly enjoyed my time at DeepMind but, unfortunately, I didn’t feel like I was able to have much climate impact in my role.
In 2018, it started to dawn on me that it’s crucial for multiple organisations to work together in order to maximise climate impact. But, one major stumbling block when trying to build these collaborations is that all the partners want to own all the intellectual property. So, maybe one way to maximise climate impact would be to deliberately not own any IP; and just release as much code and data as possible. So, I’m very excited to be building a non-profit research lab, totally focused on building things which reduce emissions ASAP.