Experience

Reviewers needed a way to check the numbers in AI-generated business reports. I designed and built a validator with AI coding tools that lets them inspect the evidence behind a figure.

It extracts claims from report text and tables, checks defined metrics against tested data queries, and separates those results from AI-assisted checks. If a definition is missing, the reviewer can see that the claim still needs attention.

The team uses the validator internally to review reports. It gives people a repeatable way to check the output and see where their judgment is still needed.

As I validated reports, I saw that fixing individual issues would only get us so far. Business and engineering needed an agreed way to define the correct result and test for it.

I analyzed the patterns and brought them to my director. I made the case for business-level validation alongside code testing, so engineering could check that the output meant what the business intended.

The evidence helped move the discussion toward stronger validation and a possible refactor. I now contribute to that strategy work; the refactor is not a completed result.

After my summer internship, I saw a gap where my AI skills could help. I connected that need to the work I had already delivered, pitched senior leadership and brought stakeholders together around an extension.

Then I had to prove it was worth doing. I joined a team of senior managers and directors, built tools they could use, and followed through on the problems they trusted me to investigate.

I earned my organization’s first remote internship extension outside software engineering. Within months, leaders were asking for my perspective on strategy as well as relying on my tools and verification work.

Finding a problem was only part of the job. I built a shared issue board, a ticket-linked handoff page and a retest workflow so the team could track what needed to change.

Each issue records the observed output and expected behavior. After engineering makes a fix, I compare the previous output, expected result and revised output against the same criteria.

The team tracked 118 report issues, most found through my checks. I retested 45 engineering fixes, returned 18 for further work and caught 2 deployment blockers. Engineering adopted my verification as a business sign-off step.

AI-assisted work can lose context when a new chat starts. I designed Workspace Kit to keep project evidence, decisions and commitments organized so the next session can pick up the work.

The kit gives each project its own records and a short briefing, with incoming updates kept separate from settled decisions. I packaged the approach for teammates and use separate Claude and Codex implementations.

Across my work, I built 6 workflow tools and packages used by the team and engineering. Workspace Kit is one part of that broader tool-building effort.

I am developing a way to evaluate whether AI-assisted supplier emails explain the important changes clearly and accurately.

Using a senior analyst’s rubric, I am building a test set and automated grader, then comparing its judgments with expert assessments.

Built the initial grader and identified where its judgments diverged from expert assessments. I’m calibrating it against those assessments before using its scores to guide pipeline changes.

I built a reusable recommendation engine from 8.9 million records and 45 pages of scoring rules, then directed an app that let merchants explore recommendations and bundles without coding.

Candidates are filtered and scored as recommended, needing verification or excluded, with a reason for each decision. I preserved existing human labels and kept recommendation rules separate from the AI explanation layer.

Engineering adopted the 5,000+ recommendations as a source-of-truth dataset for its retrieval prototype. A teammate contributed bundle rules; engineering built retrieval. The app was shared for internal testing.

Customers need a useful way to explain a project before a shopping assistant can help. I explored visual project intake and led research into the questions customers encounter on competitor product pages.

I led a three-person study of 162 suggested questions across 39 pages. In separate testing with four customers, the visual intake approach was preferred. I used that feedback to think through how to collect useful context without making the intake harder to complete.

I led a seven-person team to second place in the intern innovation competition and co-presented the concept to colleagues and senior leaders.

I built an interactive workshop site and taught AI-assisted building to 65+ colleagues.

The workshop connected the tools to practical work through an interactive learning experience.

The workshop site reached 178 unique visitors in 30 days, extending the learning experience beyond the live session.

The wage concern had been raised before without winning approval. Working on the floor and knowing the local market, I believed we were asking a lot of employees while paying less than competing jobs. I built an 85+ company wage survey to test that concern.

The Regional VP was concerned that raising starting pay could leave existing employees feeling undervalued. During a business review, I spent roughly two hours working through that concern and the market evidence with him. I wanted him to understand what employees were comparing us against.

The Regional VP approved increases shortly afterward. The original store proposal was $1.50 an hour; broader district increases affected about 100 people. By the following year, location turnover had fallen from over 200% to about 75%.

I served as acting general manager 15+ times at a $50M location and hired 15+ employees through 50+ interviews.

My responsibilities combined day-to-day operations, staffing decisions and management coverage alongside my studies.

Hired 15+ employees and provided repeated general-manager coverage at a $50M operation.

I identified a food-service workflow that AI could improve and brought a concrete proposal to the CTO.

I pitched the food-service workflow proposal directly to the CTO, estimating up to two hours saved per manager each week.

The company later piloted the concept. The proposal projected up to two hours saved per manager each week.

I founded AP Visions at 16 and brought on Craft Distillers, Caddell & Williams, and Energy Recruiters as clients. I was responsible for understanding each business and turning that into a marketing approach I could deliver.

I developed campaign ideas, produced multimedia content and adapted the work to each client’s audience and brand.

The agency grew to three retained clients and $2,000 in monthly revenue at its peak.

I managed Google Ads and SEO work and tested ways to improve visibility and engagement.

I used campaign performance, A/B testing and direct client feedback to decide what to adjust.

Campaign work delivered 150% visibility growth, 400% engagement growth and 20,000+ monthly impressions.

As a Digital Marketing Assistant with a data analytics focus, I tracked social-media and SEO performance for campus recreation.

I delivered weekly analytics presentations to the recreation leadership team, translating channel performance into implications for content.

Leadership received a recurring view of what audiences responded to and what to consider next.

I prepared an analysis of hockey-championship coverage and supported student survey participation.

I examined the performance of the coverage to inform future content strategy.

Hockey-championship coverage recorded a 400%+ engagement increase. Across the semester, I grew the student following to 3,000 and secured 2,000 survey submissions.

I co-lead the University of Tennessee, Knoxville program as part of the inaugural OpenAI Student Collective cohort.

I help shape opportunities for students to build with ChatGPT and Codex, combining structured learning with feedback on their own ideas.

Co-designed a campus program around practical building and feedback on students’ own ideas.

We co-designed a semester plan with 4 workshops, 8 Studio Hours and a student showcase.

Workshops introduce the tools; open building time lets students bring a problem, try an approach and get feedback.

Designed a semester pathway from tool introductions to supported building sessions and a student showcase.

I am pursuing a B.S. in Applied Artificial Intelligence, with expected graduation in May 2027.

I connect technical coursework with product development, analytics and collaborative projects.

Expected graduation is May 2027. I am also part of the Chancellor’s Honors Program and Alpha Kappa Psi.

In our honors-course project, Align9, our team researched how to help people find relevant local services. We shared user research, feedback and product direction. I built the application.

The navigator uses structured county and service records, with a staff maintenance workflow and a distinction between published and held resources.

The work is a supervised pilot, with service and handoff validation ongoing.

For my AI 201 class, our team created Beyond Agreement to explain AI sycophancy. In a separate traffic-fatality analysis, I modeled the data and developed the insights and visualizations.

For the traffic analysis, my teammate sourced and cleaned the dataset. I handled modeling, insights and visualizations.

The case studies show the actual ownership, methods and limitations of each project.

My credentials include Home Depot AI Specialist, Programming with JavaScript from Meta, and Fundamentals of digital marketing from Google.

These complement my degree work across AI, programming and marketing.

I apply that mix through the projects and work examples throughout this portfolio.

I completed coursework at Wake Forest before transferring to Tennessee for Applied AI.

My studies explored the connection between technical work and business decisions.

I studied at Wake Forest from August 2024 to July 2025 before transferring to Tennessee.

I participated in the CHARGE Leadership Program and Alpha Kappa Psi.

Alongside my studies, I gained practical marketing experience through my separate campus recreation role.

My separate campus recreation role put marketing analytics into practice; that work is detailed in the marketing section.