Designed and implemented a customer segmentation product in Python using clustering algorithms and exposed via a FastAPI and deployed using AWS API Gateway and Lambda, resulting in 20% increase in customer retention for our customers.Pivoted our intelligence solution to hosting business logic in a Plotly application on ECS and embedding client side in a React web app resulting in weekly instead of monthly updates of our data products allowing us to grow from 0 to 20 customers and establish market fit.Managed the creation of a cutting-edge ad budget optimization tool, forecasting clients' anticipated revenue based on their marketing expenditure. Delivered a notable 12% boost in client revenue through successful implementation.Constructed a robust CI/CD module using Terraform in conjunction with CodePipeline, implementing branching strategies, mandatory code reviews, and seamless automated deployments upon successful merging.

I am aiding in the research of trading biases in financial markets. Biases such as memory and confidence play a key role in the decisions that lead to an investment decision. As part of my work I am assisting Professor Paul Smeets and Professor Katrin Godker. My key responsibilities are data processing, numerical analysis, and statistical analysis of investor transactional data.

Built a real-time bidding deep learning algorithm with Pytorch, optimizing for speed and accurate prediction. Delivered a model that beats the incumbent production model by 8% in prediction power with an acceptable inference time.

I completed a 9 month internship in the product department of rapidly expanding team. The team and I were tasked with designing and troubleshooting the transactional management and payment systems. We analysed potential for operational improvements and automation.We were successful in implementing and deploying a new payment process and foreign exchange system. I contributed to design decisions, task documentation, and played a key role in software testing and troubleshooting.