Build Fellowship: Reinforcement Learning Controller for Lunar Lander

As a Build Student Consultant (Remote, Feb 2025 - Apr 2025), I am engaged in a cutting-edge project to develop a Reinforcement Learning (RL) controller for a simulated Lunar Lander. This fellowship involves:

• Conducting Q-learning experiments using PyTorch and OpenAI Gymnasium within the LunarLander-v3 environment.
• Training Deep Q-Network (DQN) agents over 30,000+ episodes, meticulously optimizing ε-greedy strategies and batch scheduling for peak performance.
• Developing and validating theories on agent behavior, contrasting collaborative versus independent learning paradigms.
• Visualizing training outcomes through comprehensive reward plots and dynamic training GIFs, leveraging Colab GPU resources for computational efficiency.

This program offers a comprehensive learning path, from understanding RL basics and neural networks to implementing advanced DQN techniques. Learn more on the official project page. Explore the Colab Notebooks: LunarLander Colab Notebook | MountainCar Colab Notebook.

MuniBuddy: Real-Time Transit Tracking Backend (Capstone Project)

As a Backend Developer for the MuniBuddy Capstone Project (City College of San Francisco, Jan 2024 - May 2024), I spearheaded the development of a real-time public transit tracking system. Key contributions include:

• Engineered the backend using Python, FastAPI, PostgreSQL, SQLAlchemy, Redis, and Docker for robust and scalable performance.
• Integrated a React + Vite frontend with backend APIs to display live transit arrival times and vehicle positions on an interactive map.
• Implemented map-based stop detection, efficient location-based caching, and secured HTTPS deployment using Caddy on a DigitalOcean droplet with a custom domain.
• Designed and built support for a Flutter Android application, including backend optimizations for mobile performance and a planned Firebase push notification integration.
• Architected the backend to seamlessly support both web and mobile clients, with a strong focus on scalability and low-latency data delivery.

MuniBuddy Project Demo | MuniBuddy Project Swagger UI | View MuniBuddy on GitHub.

Salesforce Donor Management Application (Internship Project)

During my Salesforce Development Internship (Aug 2023 - Nov 2023), I led a team of 10 interns in an Agile environment to develop a comprehensive donor management system on the Salesforce platform. My responsibilities and achievements encompassed:

• Languages and Tools: Apex, LWC, JavaScript, MuleSoft, JSON, Trello, HTML.
• Integrated Give Lively for secure and efficient online donation processing, streamlining financial transactions.
• Automated real-time donation tracking workflows, ensuring accurate contribution records and instant acknowledgments for donors.
• Enhanced the user interface (UX) for seamless navigation, optimizing data entry processes and tracking capabilities for organizational staff.
• Developed custom dashboards and reports to analyze donation trends, donor demographics, and overall financial performance, providing actionable insights.
• Significantly enhanced transparency and efficiency in donor management, directly supporting the organization's mission and community engagement efforts.

This project was recognized with the City College Of San Francisco Fall 2024 Dean's Honor List for leading the team in this successful development.

Open Energy Dashboard (OED) - Open Source Contribution (SWE Micro-Internship)

Through a CTI / CodeDay Micro-Internship, I contributed to the Open Energy Dashboard, an open-source project. My key contributions involved:

• Implemented and validated the B17 test case in Mocha/Chai to ensure accurate energy reading from flow meters.
• Resolved PostgreSQL authentication issues within Docker containers, facilitating smooth CI test execution in GitHub Codespaces.
• Refactored `readingsBarMeterFlow.js` to improve test consistency and overall code structure.
• Debugged errors related to missing CSV files and adjusted test logic for 75 days of 15-minute interval data.
• Validated the Readings API for bar chart generation, daily bar widths, and correct Electric → kW unit conversions.

GitHub Issue #962 | View My Merged Pull Request.

Encoder/Decoder Application (Bit Manipulation)

This web-based Encoder/Decoder application demonstrates advanced bit manipulation techniques for encoding and decoding text. Tailored for uppercase strings using a custom scramble map, and standard strings via binary transformations, it showcases cross-language proficiency with implementations in both JavaScript and Python. Key techniques include efficient binary-to-decimal/ASCII conversions, Python list comprehensions, and advanced JavaScript DOM manipulation.

Try the Encoder/Decoder Web App
View Python Solution on GitHub

TOYTOPIA - Responsive Web Design with Bootstrap

If you're interested in exploring a project where I utilized Bootstrap to create a responsive and visually appealing webpage, you can check out TOYTOPIA. This project highlights my ability to efficiently implement modern frameworks for clean, user-friendly designs.

Visit TOYTOPIA