I’m currently employed at Busuu, a popular language learning app. I’ve built many things there, the most interesting of which is a phonetic speech recognition system that transcribes user’s speech into IPA symbols so they can receive detailed feedback on their pronunciation. The model utilises unsupervised learning to squeeze the most performance out of the relatively small dataset available on this problem.
Before that I built a model to classify URLs into phishing/safe at Wandera and earlier still I studied physics at University College London.
There are many fields of machine learning that I’d love to work with in a professional capacity:
- Reinforcement learning
- Unsupervised learning (especially in combination with the above)
- Speech recognition/speaker identification
- Creative/generative models