CS @ Gannon University '27 · Data & Backend Engineering

I build data and backend systems end to end, and I check the numbers before I report them.

Currently an undergraduate research assistant working on empirical software engineering, and the student developer behind Gannon STEM Center's internal tooling. Graduating May 2027 — open to full-time software engineering, data analyst, and AI/ML engineering roles starting mid-2027.

Olist Delivery Performance Analysis · SQL Server 2022 · T-SQL · Power BI · DAX

Traced 87% of lost delivery days to the courier leg rather than sellers across 99,441 orders. Two bugs caught in T-SQL validation first — one had silently misclassified 1,292 orders. 1.55M rows through a staging → cleaning → star-schema pipeline that logs rejects instead of dropping them.

AdventureWorks Sales & Profitability Analytics · Power BI · DAX · Dimensional Modeling

Three linked dashboards over $29.4M of internet sales. Separating revenue from margin corrected a misleading read: bikes carry 96% of revenue at 40.6% margin, accessories return 62.6% on a smaller base.

NLP Sentiment Analysis & Opinion Mining · PyTorch · scikit-learn · Streamlit · FastAPI

Benchmarked 9 models over 19,562 Amazon reviews. Macro F1 exposed what accuracy was hiding on an 83% positive class — weak neutral-class recall the weighted metrics concealed.

Task Management API · NestJS · TypeScript · PostgreSQL · Redis

Domain-driven backend — 23K lines, 147 modules, four bounded domains, 54 endpoints. Task workflows modeled as an explicit state machine with an event-driven webhook layer.

Undergraduate Research Assistant, Gannon University · Jan 2026 – Present Faculty-supervised research under Dr. Md Tajmilur Rahman — co-authoring "Understanding Characteristics of Browser Bugs: A Case Study on Mozilla Firefox."

Analysis of 247,959 Firefox Bugzilla reports: severity and priority behave as largely independent dimensions; component predicts priority but barely predicts severity; semantic text clusters predict both better than the official component taxonomy does.

The result I'm proudest of is one we didn't publish — an association that looked real at Cramér's V 0.285 collapsed to zero once corrected for sample sparsity, so we reported it as an artifact instead of a finding.

Also: T-SQL · DAX · Power Query (M) · star schema & dimensional modeling · ETL · CTEs & window functions · Prisma · TypeORM · LangChain · RAG & vector search · GitHub Actions