• Implemented 15+ client-driven features on a Java-based platform processing 30K+ monthly transactions (e.g., securities, ACH/wires), integrating AlphaFrontier software and utilizing Kafka for efficient data streaming and message consumption

• Investigated and resolved 15+ system defects through SQL debugging and GWT log analysis, increasing reliability and minimizing production risks

• Tested features across 10+ Agile sprints and collaborated with QA to validate system enhancements in development and UAT environments

• Engineered 5+ AI/ML applications using Next.js, Flask, OpenAI, and vector databases, following Agile methodologies and the full SDLC process, culminating in a flagship project achieving over 1,000 waitlisted users

• Spearheaded the development of a multimodal customer complaint analysis platform for a fintech startup; integrated Google Cloud APIs and OpenAI, and leveraged Postman for efficient API testing and validation, resulting in a significant reduction in processing times for customer inquiries

• Launched an interactive application focused on holistic health, integrating a RAG pipeline to source data from 40+ YouTube videos by reputable experts and using the data to train 5 AI chatbots to provide personalized guidance

• Transformed a legacy SQL stored procedure into a streamlined Python script, reducing execution time by 90% through logic and query optimizations, improving the efficiency of financial reviews

• Addressed critical production issues on server-side via SQL queries and HighJump/Koerber WMS, leading to the effective management of 300K+ inventory items per day

• Conducted end-to-end testing prior to deployment and wrote clear design documentation for new features, streamlining workflows for 60+ analysts and developers

• Developed a robust crime mapping tool using Python, Pandas, GeoPandas, and Matplotlib to analyze 7M+ crime records and generate geographic heatmaps highlighting high-risk areas

• Cleaned and processed millions of lines of crime data using Python and CSV files, improving data accuracy by 40%