M.S. Data Science, Seattle University · second year
Duy Nguyen
I find out whether a machine-learning result is real, build the data it depends on, and explain it to the people who decide.
Background: an AI-generated composite painting of the Seattle skyline and the Seattle University campus, animated by AI — not a photograph or footage of either.
A paper, a database, an award
Three outcomes this year. Everything else is on the résumé.
- PSB 2027
First-author paper, accepted and selected for an oral presentation
Pacific Symposium on Biocomputing · January 2027 · Big Island of Hawaii
With Dr. Wenjing Yang, on retrieval-augmented mammography report generation.
Accepted, not yet published: the camera-ready revision is due 1 October 2026 and the manuscript is not public until it is submitted; it then appears in the proceedings, which are indexed in PubMed, and the oral presentation follows at the meeting in January 2027.
- SU CNS
Research Data Engineer, Computational Neuroscience Research Group
Prof. Brian Fischer · Seattle University · NIH CRCNS-funded · March 2026 – present
I designed the schema and wrote the Python ETL that loads XDPHYS, MATLAB, and CSV files into one SQLite database. Lab students now run neuron-level queries themselves; the database is available on request.
195 neurons1,325 recording passes30,147 raw files
This work is ongoing, so its counts move; every figure here was verified against the lab database on 2 September 2026 and is re-derivable from it.
- CAUSE 2026
Graduate Division winner, CAUSE data-story contest
CAUSE Student Data Scrollytelling Contest · nine acts
Judged blind on narrative, data interpretation, scrollytelling, and visual design. Seattle University Graduate Studies interviewed me for a story on the win.
Also live: mavterras.com, my brother’s construction company. I turned two founders’ requirements into a scope, wrote the code with AI, and I deployed it and operate it in production: every customer brief and photo set is saved on arrival, sent to the company, and given an AI-drafted pre-call brief. It runs on Vercel, Neon Postgres and Cloudflare R2, the Anthropic API calls sit behind a demo-mode lock that defaults on, and when that pipeline fails or the database is cold, that is mine. Next, and mine to design and build: the company’s own job and agent system for internal operations, built in-house instead of another third-party service, so every lead is qualified and the homework done before anyone drives to a site.
Paste a job description or pick a role
You get a short brief: which of your requirements I can evidence, which only partly, and which I cannot, each tied to a specific piece of work. It reads the same record this page renders from, and where there is no evidence it says so.
Or ask a question
Hiring for 2027?
I graduate in June 2027, after all 45 credits and the two-quarter capstone. I am open to both: an internship in the summer of 2027, and new-graduate research, data-science and machine-learning-engineering roles starting mid-2027.
The résumé has the timeline; the fit brief above has the evidence.