Built an LSTM-based financial time-series model to forecast NVIDIA's stock price, achieving strong out-of-sample accuracy with expanding-window validation. Demonstrates expertise in risk modeling, portfolio forecasting, and algorithmic trading applications.

Developed a 124M-parameter GPT-2 model from scratch with GPU acceleration, optimized for limited hardware. Relevant to regulatory compliance, financial sentiment analysis, and natural language processing in FinTech.

Designed a domain-specific query language for CSV databases with SQL-like syntax and a VS Code extension for syntax highlighting. Mirrors real-world database engine development and strengthens expertise in systems-level programming and data infrastructure.

Built a safety-critical runway recalculation tool with declared-distance algorithms and obstacle visualization, using Agile SCRUM. Demonstrates high-stakes software engineering and transferable skills for mission-critical FinTech systems (e.g., payments, compliance).

Developed a CNN-based MRI tumor detection system with automated preprocessing and a Streamlit dashboard for clinicians. Showcases ability to deliver end-to-end ML pipelines with real-world usability, relevant to healthcare and anomaly detection use cases.

Created a Java application with a recommendation engine that generates personalized recipes using ingredient + weather data. Includes a gamified UI, demonstrating practical product design and user-focused development.