- Developing a subscription-based product that provides AI and algorithm-driven trading signals for retail investors, guaranteeing a minimum of 10% monthly yield based on backtest results.- Joined as an intern and transitioned to a full-time Machine Learning Engineer role, tasked with transforming the monolithic ML pipeline into a scalable, microservice-based architecture.- Led the redesign of the ML pipeline into 15 microservices, deployed via Kubernetes and Docker on Exoscale Cloud, resulting in a 30% improvement in prediction accuracy and a significant reduction in cloud costs through efficient resource allocation.- Implemented features such as hyperparameter tuning using the Optuna framework, advanced error correction techniques for time series models, and additional model architectures, contributing to the system's scalability and accuracy.- Developed an MLOps tool automating model development for different assets and cases, significantly streamlining processes for the team.- Reduced operational costs by implementing dynamic creation and deletion of GPU node pools during model training, ensuring GPU machines were used only when needed.

• Developing a pipeline that can extract data from Meta Marketing Insights API on a daily, weekly, and monthly basis and store in a MySQL database.• Developing an API with Flask framework for providing relevant statistics for marketing analytics like cost per action, cost per click, conversions, click-through rates etc. • To visualize the statistics, created an interactive web dashboard with using Bootstrap and JavaScript that displays KPIs and graphics.