GitHub: https://github.com/WBGu/NFL_Game_Prediction?tab=readme-ov-file

• Built an end-to-end NFL game outcome prediction system using XGBoost with time-aware validation, Optuna hyperparameter tuning, SHAP-based explainability, and probability calibration, producing interpretable and well-calibrated win probability estimates. 

• Implemented a custom Elo rating system and integrated it into a feature-engineered NFL prediction pipeline while preserving strict separation between labels and engineered features. 

• Trained and tuned an XGBoost model using time-aware cross-validation and Optuna hyperparameter optimization, achieving improved log-loss and ROC-AUC over Elo and logistic regression baselines. 

GitHub: https://github.com/WBGu/BlackScholesModel/tree/master

• Designed and implemented a production-quality C++17 Black–Scholes pricing library featuring analytical pricing, full Greeks, multithreaded Monte Carlo simulation, and implied volatility solving.

• Built a robust implied volatility solver using Newton–Raphson with Vega and a bisection fallback to guarantee numerical convergence.

• Implemented a multithreaded Monte Carlo pricer with near-linear scaling up to physical core count; isolated RNG per worker to avoid contention.

• Benchmarked and profiled pricing throughput using chrono-based microbenchmarks and CPU profiling tools, identifying compute-bound and RNG-bound hotspots.

GitHub: https://github.com/WBGu/OrderBook

• Implemented a real-time order book engine using max/min heap data structures with price-time priority and O(log n) insert/remove using C++

• Built a matching engine supporting limit/market orders, cancellation, and FIFO execution within price levels to simulate real exchange behavior.

• Developed a Qt-based GUI with live depth charts that aggregate liquidity by price level and visualize cumulative bid/ask volume in real time.

Github: https://github.com/WBGu/InventoryMangement

• Developed a desktop (GUI) for inventory management using Python, Tkinter, and the tksheet library.

• Engineered real-time data validation and conditional UI formatting that cross-references user input against a local database to prevent overselling and eliminate invalid item entries.

• Integrated Git subprocesses and native Windows Batch scripts directly into the application, enabling automated, one-click synchronization between the local data store and a remote repository.

• Secured data integrity by integrating automated Git version control commands via custom batch and bash scripts, ensuring the master JSON inventory file is consistently backed up to a remote server.

Github: https://github.com/WBGu/ItemLookUp

Webapp: https://itemlookup-kexfjmfmtap9gms7nxsj7b.streamlit.app/

• Developed a web-based internal search tool using Python and Streamlit to query and retrieve images from Google Drive directories using Google Drive API.

• Engineered a recursive folder-crawling algorithm to overcome native API limitations, enabling deep, multi-level searches across nested subfolders.

• Transitioned from an Android APK build to a Streamlit web app, making the tool instantly usable on iOS, Android, and Desktop without installation. 

GitHub: https://github.com/WBGu/RLearning-Trading-Bot

• Investigated using a random forest (boostrapped ensemble of random tree learners) for deciding stock holdings during a given time window, using trading signals as features

GitHub: https://github.com/WBGu/CosmicCrusaders

Netify Host: https://cosmic-crusaders.netlify.app/

• Developed a dynamic enemy wave spawning system leveraging a hashmap to manage wave difficulty and implemented random generation to introduce 10 unique enemy types with varying behaviors, including patterned movement and AI-driven player tracking. 

• Designed and implemented the core player attribute system (health, max health, boost, max boost) within an object, employing an array for structured data storage and streamlined access through Object-Oriented Programming. 

• Organized and led frequent team meetings to monitor progress, address challenges, and ensure completion of tasks across a team of three

Github: https://github.com/WBGu/WhatNot_PackingSlip

• Developed a C++ and Qt GUI application that streamlined Whatnot's shipping process by transforming raw CSV data into condensed printable PDF packing slips, reducing handling time by 148% and significantly minimizing errors for sellers.

• Engineered a robust CSV parsing method with custom logic to accurately process fields containing irregular formats, ensuring data integrity by leveraging Regex and delimiters.

• Implemented Object-Oriented Programming (OOP) principles to manage buyer information and product data, utilizing various data structures for dynamic category management.

• Integrated a merge sort algorithm to organize and sort item within each buyer's various product categories while maintaining clarity and usability of order details.

• Categorized items using string manipulation techniques and hash table lookups to group and display the diverse inventory.