CS + BUSINESS @ MSU
I build data products, full-stack platforms, and applied AI tools that turn messy workflows into useful software.
About
Music content creator
Work
Technical and leadership roles across data systems, consulting, student organizations, and residential life.
Tesla
Built an entity-resolution pipeline for a 35K-account installer network and replaced manual Excel workflows with a Power Apps CRM for 70+ Account Managers and Sales Ops users. Developed GenAI tools supporting 100+ users across 20 workflows and executive-facing solar analytics.
EY
Supported a $13M enterprise migration by validating 10K+ records, automating SQL Server checks, assisting Azure Blob Storage migration, and triaging defects in Azure DevOps.
SpartaHack
Co-lead planning and execution for MSU's 600+ attendee student-run hackathon, overseeing Tech, Marketing, and Logistics while building the live platform for participant and organizer workflows.
International Students Association
Led ISA's Next.js and Supabase platform from planning through production release, bringing its public website, event registration, affiliate attendance, funding, and staff workflows together.
The Global Career Accelerator
Analyzed 30M+ engagement records, evaluated Grammy.com KPIs, designed A/B experiments, and evaluated engagement differences at 95% confidence. Built SQL/Tableau dashboards and modeled regional energy readiness for Intel-sponsored data center analysis.
Michigan State University
Support 80+ residents through conflict management, community programming, crisis response, and day-to-day community support.
Education
The academic foundation behind my work in software, data, and business systems.
Michigan State University
Maintaining a 3.95 GPA while studying computer science, business, artificial intelligence, big data analysis, cloud development, and data structures.
Selected Builds
Selected work across machine learning, backend systems, internal tools, and interactive software.
A Streamlit/PostgreSQL market intelligence platform for company scoring, valuation, catalysts, and investment thesis tracking. Its AI companion, Argus Friend Brief, turns company research into plain-language, evidence-linked briefs using local Gemma.
Read case studyAI infrastructure research was scattered across market data, SEC filings, news, catalysts, watchlists, and investment notes, making it difficult to evaluate companies and themes consistently.
Built scheduled ingestion pipelines and a 33-table PostgreSQL data model for company research, valuation, thesis tracking, and catalyst analysis. The AI companion runs Gemma locally through Ollama, validates citation IDs against a bounded evidence catalog, and rejects invalid responses after one repair attempt.
Created a research platform tracking 53 companies, used daily by my dad and refined through his feedback. It combines explainable scoring and 5-, 20-, and 60-day backtests with plain-language research briefs, supported by behavioral tests and PostgreSQL-backed CI.
A Next.js and Convex platform for SpartaHack's student-run events, connecting participant workflows with organizer operations.
Read case studyA student-run hackathon designed for 600+ attendees needs more than a landing page: participants and organizers need reliable registration, scheduling, judging, check-in, communication, and administrative workflows.
Independently built participant applications, teams, project submissions, check-in, judging, and organizer workflows. Added backend event isolation, department-scoped staff permissions, and durable communication jobs with retries and duplicate-safe recovery.
Published the live platform for participant registration, check-in, and organizer operations ahead of SpartaHack 12.
A production Next.js and Supabase platform for public events, registration, affiliate operations, attendance, funding requests, and staff administration.
Read case studyISA needs a more structured way to manage affiliate organizations, track participation, and handle funding request workflows.
Built public and staff workflows with server-enforced permissions, organization isolation, conflict-of-interest rules, and audit records. Tested duplicate-safe operations and automated releases and PostgreSQL migrations through CI/CD.
Launched msuisa.org, bringing ISA's event registration, affiliate attendance, funding requests, and staff administration into one platform with protected access and traceable operations.
An MHacks-winning Streamlit platform for optimal nitrogen prediction, crop yield projection, and farmer-facing soil resources.
Read case studyFarmers need more personalized guidance around soil health, fertilizer use, and crop selection, especially when fertilizer costs and overuse create practical and environmental tradeoffs.
Worked on the Streamlit frontend and ML models, using crop and soil datasets with temperature, humidity, pH, rainfall, crop weight, moisture, crop type, and harvested area features.
Won Best Sustainability Track at MHacks 2024 with a 24-hour prototype for nitrogen recommendations, crop yield comparison, and farmer resource access.
A Dockerized Flask microservices backend for a food marketplace, with REST APIs, JWT authentication, SQLite, and service orchestration.
Read case studyA distributed marketplace backend needs clear service boundaries for authentication, orders, search, logging, and product management.
Built 5 Flask microservices with REST endpoints, JWT-based authentication, SQLite persistence, Docker Compose networking, and separated transaction workflows.
Created a cloud-deployment-ready backend foundation that strengthened service orchestration, authentication, and reusable architecture patterns.
A PyTorch computer vision project that fine-tuned AlexNet to classify a 17-class celebrity face dataset.
Read case studyImage classification models need to distinguish between visually similar faces while handling variation in lighting, pose, and image quality.
Fine-tuned an AlexNet-based model using transfer learning, data augmentation, dropout regularization, k-fold cross-validation, and confusion-matrix error analysis.
Achieved 57.35% classification accuracy across 17 celebrity classes, using confusion-matrix analysis to inspect errors.
Skills
Tools and skills grouped by how I use them.
Python, SQL, PostgreSQL, SQLAlchemy, ETL Pipelines, Data Validation, Entity Resolution, GitHub Actions
Pandas, NumPy, Power BI, Tableau, Excel, KPI Design, A/B Testing, Reporting Systems
GenAI Workflows, MCP, Local LLMs, scikit-learn, PyTorch, Random Forest, Transfer Learning, Computer Vision, Model Evaluation
TypeScript / JavaScript, Next.js, React, Supabase, Convex, Server-side Authorization, C/C++, Azure, Docker, REST APIs, Flask, CI/CD, Git
Contact
Reach out about software, data, AI projects, sustainability, internships, or collaborations.
Thanks! Your message is on its way. I’ll get back to you soon.