I build LLM systems and applied ML, from research pipelines that hold up to peer review to products people actually use. Recent Computer Software Engineering grad, based in Chicago.
An end-to-end LLM pipeline analyzing an electrical-construction media corpus: data collection, GPT-based classification, human expert annotation, and rigorous evaluation, plus a deployed retrieval-augmented chatbot. Flagship research, journal manuscript in progress.
Browser games that turn ElectriAI's research into hands-on learning, from arcade quizzes to Arc Descent, a full 3D action dungeon crawler where questions are the loot economy. Every game is a dependency-light browser build with a question-driven economy, CSV/JSON question upload, and localStorage progress, mapping to the project's 10-class electrical-construction schema.
An Electron desktop control center built to run an entire AI-assisted workflow from one window: parallel Claude Code sessions in embedded PTY terminal grids with saved workspace layouts, pop-out panes, and broadcast typing; a full file browser over an Obsidian vault with markdown editing, wikilink navigation, search, and share-to-phone; a 3D knowledge graph; live token-usage analytics; and a gamification layer (XP, levels, quests, pet companion, shop) that turns daily work into progression. The live demo is a faithful in-browser simulation on mock data.
Ensemble-based uncertainty quantification for red-tide (Karenia brevis) bloom prediction on Florida's Peace River. Four UQ methods (Bootstrap, GLUE, EnKF from scratch, LPU) build 200-member ensembles that propagate input uncertainty through an ML classifier for calibrated probabilistic forecasts. Grew from notebooks into a tested, packaged Python library.
A multi-service LLM platform that grades uploaded pitch decks across six categories and persists a history view. React frontend, FastAPI/PostgreSQL persistence service, and a Node/TypeScript grading service with structured JSON-schema output, adaptive retry, and non-blocking DB writes so the API stays up if the database is down.
LLM programming tutor that teaches through Socratic hints, not solutions. Modular system-prompt architecture adapts to skill level.
Python · FastAPI · React · PostgreSQL14-model scikit-learn benchmark on 32,416 loans. Full preprocessing, tuning, and evaluation pipeline; 0.95 test ROC-AUC.
Python · scikit-learn · pandasNormalized five-table schema modeling fighters, fights, and stats. Aggregation queries for win/loss ratios and outcomes by class.
PostgreSQL · SQLEight Selenium + TestNG UI test classes driving a live browser through login, posting, search, and recovery paths.
Java · Selenium · TestNG · MavenFull Tetris recreation from scratch: rendering, collision, line-clearing, and scoring in C++ with SFML.
C++ · SFMLArduino security sensor written in AVR assembly on the ATmega328p: ultrasonic ranging and alarm logic.
Assembly · ATmega328pI'm a software engineer who leans hard into AI/ML. My work spans the full arc: collecting and cleaning messy real-world data, building and evaluating LLM and ML systems, and shipping them as real products.
My flagship is ElectriAI, an independent research project I built end to end and am publishing, advised at FGCU and funded by the Head Start Grant. Alongside it I've shipped full-stack LLM apps and a tested scientific-computing package, with a research engineer's habits: reproducible, evaluated, and honest about what the numbers say.
Comfortable across the stack, from PostgreSQL and REST APIs to systems-level C++ and assembly, with strong fundamentals in data structures, algorithms, and operating systems.