• Built a DRL environment in Python using TensorFlow, PyTorch, and OpenAI Gym for real-time trading in markets
• Developed and trained a DRL agent leveraging sequence models (LSTMs) for market data analysis and trading execution
• Led a workshop on the principles of machine learning in Python and its applications in predicting the markets

• Designed and developed an inventory logger and profit and loss calculator to streamline accounting workflows, employing Pandas, NumPy, and Matplotlib, libraries in Python
• Showcased the proof-of-concept to support financial planning and resource management through data-driven insights, able to increase accounting team efficiency by ~30% and improve forecasting accuracy

• Piloted a research team to assess the strengths and risks of the Chinese financial market, delivering actionable insights on market resilience and stability to senior management
• Conducted data filtering and utilized MATLAB’s experiment mode to process and model extensive financial datasets, analyzing market volatility and structural patterns

• Organized and managed workshops with the QFCRA; covering ethics, finance, corporate law, and data science
• Connected exceptional students with internships, then guiding them through tasks and proposal development

• Developed a web app with Gemini API for sentiment analysis of pictures, streamlining content creation on Instagram
• Developed a React frontend with an Node.js backend, optimizing images with Multer cutting processing time by %40

• Crafted a robust class structure with inheritance and encapsulation to model piece and board interaction
• Employed arrays, array lists, 2D arrays, and recursive functions to handle game logic and piece movement

• Developed a stock indicator that leverages data from Google Finance to calculate fair stock values, integrating metrics such as price, daily % change, 52-week high/low, and P/E ratio for initial valuation assessments
• Enhanced the model to calculate stock momentum, comparing it to the S&P 500 to highlight market trends
• From March 2021 to January 2024, the model returned 56% compared to the S&P 500’s 22%

• Collected and analyzed elevator traffic data to understand patterns and peak usage times
• Implemented a dynamic algorithm to optimally manage traffic based on the time-of-day, reducing average wait times
by ~50%; collaborated with engineers to integrate the solution into Mitsubishi’s proprietary elevator software