Deep Learning

Extracted information from legacy contracts, built taxonomoy to categorise them and linked contracts hierarchically

Explored and created a web-hosted demo for a zero-shot deepfake audio generation model to generate audio from text using embeddings of short reference audio clips

Under Review

Explored an ensemble of deep learning and evolutionary strategies as an RL alternative in the context of an investment problem

Explored distributing model components across different devices to enable running large models on edge with multiple smaller devices

Explored concept drift and built an sklearn-like toolkit to detect drift in image, text or audio data using multiple drift detection methods for monitoring and extending life of production models

Analysed and predicted the behaviour of visitors in NYC's Bryant Park, building a digital twin spatial model for 4D simulations

Extracted HTML UI elements from wireframe drawings of websites and website screenshots using advanced image pre-processing and confidence cutoff variation. Achieved $2^{nd}$ rank in the hackathon.

Participated in the Twitter Bias Hackathon and highlighted bias instances in Twitter's Saliency model used to crop images building a case study with Tokyo Olympics images as example. Achieved $9^{th}$ rank in the hackathon.

Predicted demand using sales figures, mobility data and several demographic and footfall variables