Hi there! I'm James, a Machine Learning Engineer at Bitfount!
Building out our Federated Learning platform from the machine learning/Python side; producing the framework and libraries needed to provide an easy-to-use federation system to allow remote training of ML models and execution of federated queries. I'm also primarily involved in the creation and application of our privacy-preserving technologies to ensure that user data is kept secret whilst still allowing state-of-the-art models to be trained on them. My day-to-day is mostly coding in Python and system architecture design.
I did a BSc in Computer Science and then followed that up with a PhD in Artificial Intelligence, focussing on Multi-Agent Systems. After the PhD I got a role working on a multi-agent project at Samsung Research UK and when that project ended moved into a team working on privacy-preserving on-device federated learning for voice recognition models. From there the role at Bitfount was a perfect fit as I was able to build on the knowledge I'd acquired whilst working on a new system from scratch!
At the moment I'm focussed on ensuring the high throughput of our messaging system to ensure that the federated learning can proceed as quickly as possible whilst also making sure that the modeller is provided updates as to how training is progressing.
Outside of work I'm a keen gamer of all kinds (video, board, and TTRPG) alongside my wife and two cats (though they're less involved in the gaming aspect!)
The culture at Bitfount is hands down one of the best I've worked in, having worked in a number of different organisations and company types throughout my PhD and afterwards: public sector, scale-up size companies, academia, large multinationals. The people are wonderful, passionate, and humble, and it's great opportunity to continue to grow and learn in an environment where there's no "office politics" or egos.
Meet Sam and find out what it's like to be a Software Engineer at Bitfount
Meet Merlin and find out what it's like to be a Software Engineer at Bitfount.