• Built a streaming dashboard to visualize the real-time funding and borrow rates, along with other relevant metrics to help identify basis trading opportunities in perpetuals markets.

• Developed an internal package to interpret and explain deep learning models using the open-source SHAP library.• Containerized the pipeline using Docker and created an Airflow DAG to run it on an EMR cluster at regular intervals.

• Enhanced existing pricing models by engineering new features, improving model performance by 25%.• Implemented batch processing and integrated Apache Kafka into various microservices to reduce the response time by 20%.

• Built a multi-task learning NLP model to make predictions about commercial clients based on scraped website data.

• Built a MobileNet transfer learning model to analyze the level of screen damage and quote an initial price.• Productionized the pipeline and wrote a microservice to serve predictions.

• Developed a Power BI dashboard to track risk factors and detect anomalies using data consisting of search keywords and traffic related metrics. • Built an internal logging and reporting platform for tracking bugs.