Bio

My work is heavily centred around finding solutions to NLP / NLU tasks that is usable in a production setting and built from noisy real-world data. I use both classical and deep learning approaches to solve these problems. Translating research papers to code and applying it on new data is something I do on a day to day basis in order to find out a good tradeoff between model performance and computational costs. The end deliverable is often a hybrid of classical NLP models and state of the art Deep Learning approaches which promises both high speed and accuracy. One of my main goals is to transform how language is being modelled at the moment with novel techniques that require fewer compute while beating the current state of the art.