Projects
AI-Powered Trading Bot
Developed an automated trading bot that analyzes stock trends using historical data and real-time news sentiment analysis. It integrates AI/ML models to predict market movements and generate trading signals.
- Libraries Used: lumibot, Alpaca API, PyTorch, Transformers
- News Sentiment Analysis: Retrieves financial news from Alpaca API, processes it through FinBERT, and assigns sentiment scores.
- Trading Signal Integration: Combines stock data with sentiment analysis to generate buy/sell signals dynamically.
- ML Model: Uses ProsusAI/FinBERT on CUDA or CPU to refine trading strategies.
Digital Audio Signal Processing System
Designed a hardware-based Digital Audio Signal Processing System in SystemVerilog, computing and outputting the amplitude of stereo audio signals (left & right channels).
- Finite State Machine (FSM): Controls channel selection and processing workflow.
- Audio Signal Processing: Accumulates, scales, shifts, and averages audio data for smooth amplitude computation.
- Amplitude Calculation: Converts signed audio data to absolute values and represents magnitude in thermometric format.
- Implementation: Uses Quartus Prime, ModelSim, and FPGA synthesis for real-time operation.
Flappy Bird Clone
Developed a **Flappy Bird Clone** in **Java (Swing & AWT)** with real-time physics, collision detection, and an interactive game loop.
- Game Framework: Built using **Java Swing & AWT** for custom rendering and game mechanics.
- Physics & Collision: Implements gravity, velocity-based movement, and bounding box collision detection.
- Dynamic Obstacle Generation: Procedurally generates pipes with randomized spacing.
- Finite State Machine (FSM): Controls game flow, score tracking, and game-over states.
- Restart Mechanism: Pressing SPACE resets the game for a seamless user experience.
Shortest Path on Campus
Developed a Java application to determine the shortest path between buildings on the UW-Madison campus using Dijkstra's algorithm.
- Graph Implementation: Created custom graph structures to model campus pathways.
- Algorithm Application: Implemented Dijkstra's algorithm for efficient shortest path computation.
- Frontend Integration: Designed an interactive interface for users to select start and end points.
- Backend Functionality: Managed data processing and path calculations.
- Testing Suite: Developed unit tests to ensure accuracy and reliability of pathfinding features.
iSongify
Developed a Java-based application for managing and searching a song collection, utilizing advanced data structures for efficient storage and retrieval.
- Data Structures: Implemented Binary Search Trees and Red-Black Trees for optimized song storage.
- Backend Development: Created robust backend systems for song data management.
- Frontend Interface: Designed an intuitive user interface for seamless interaction.
- Testing: Conducted comprehensive unit tests to ensure application reliability.
