CS student @ Portland State, graduating December 2026
Team Lead at PSU Computer Science Capstone | March 2026 - August 2026
Seven of us built an open-source harness that lets a language model reverse engineer a USB device with a vendor specific protocol. I ran the sprints, the triage and the reviews, and wrote the specs. We proved it worked by decoding a USB relay's entire command protocol, frame structure and checksum included, off live captures.
Me as a
ARI3, independent research | May 2026 - Now
ARI3 is the model system behind ARI3LLA INDEX, my public archive of style language, and I run it as a small research program. The current version reads a news headline and its feed excerpt, decides whether the item is about style, and says when it isn't sure. Every version is frozen, hashed and committed before its results are published, and every experiment is a permanent entry in the notebook. The first frozen version was right 88 percent of the time on news items it never saw during training, against 53 percent for always guessing the more common answer.
Product Designer and Sole Developer at Clique | October 2024 - April 2025
Clique is a private space for one friend group, with a group chat, a shared camera roll and a calendar. My partner had the idea and the business case. I designed it in Figma, tested it one on one with 15 people across five rounds, and rebuilt the flows they got stuck in, then built all of it myself in React Native. We placed 2nd at the UP Pilot Venture Challenge on a live demo.
Project Manager at Metro Metrics, Oregon Tech | March 2026
Five of us built a walkability scoring model for external stakeholders. I ran the schedule and the scope, and presented the findings alongside my team.
Me as a
Modeling and Simulation Intern at PLEXSYS | June 2026 - Now
I build 3D assets for simulation software, and tooling that speeds that work up.
Me as a
Product Manager at VirtualBoss | September 2025 - April 2026
My employer bought VirtualBoss, a construction management application written in FoxPro, with no plan for what came next. I studied how the program was used day to day, wrote the requirements and the migration plan, and flew out for the acquisition handoff. Because I understood both the construction side and the software side, I was the person the business owner, my employer and the lead engineer all talked to, which meant the technical decisions got made against real problems from the field.
Me as a
Our sponsor made working with AI a requirement from day one. I asked everyone on my team to write down how they worked, then put my own account next to theirs. Seven people had independently arrived at the same handful of practices, and every one of them was learned by running into the failure it prevents. The failures are in there too, including a specification that got written before the group had agreed on its shape, and the two most expensive bugs of the term, both caught by a person who went and tested a claim the tests said was already covered.
I also wrote down what changed about the job of leading. Our daily standup stopped earning its slot once nobody was blocked, and the weekly meeting turned into where the leading happened.
Read the essay