Codivus
An offline AI programming-education platform combining local language models, FastAPI, React, Ollama, and retrieval-augmented generation.
Dive deeper
The projects changed. The questions became sharper.
A closer look at how hackathons, teaching, internships, graduate study, and a growing homelab shaped the way I build: practical first, evidence-led, and curious about every layer of the system.
01 / Project evolution
From useful demos to systems with a point of view.
BenchMind
An agent benchmarking platform built at Cal Hacks, extending my interest from using models to measuring how intelligent systems behave.
STEM Forge
An accessibility-focused learning project recognized with Best AI Integration at MESA U Hacks.
CuraPoint + Scaffold
Healthcare access and structured writing tools that pushed the work beyond model capability toward workflow, clarity, and human agency.
02 / Academic transition
Three undergraduate years built breadth. Graduate school is where I go deeper.
Summa Cum Laude, Dean’s List every term, Learning Assistant for CS 101, MESA mentor, club leader, and a foundation spanning AI, security, operating systems, Linux, theory, biology, and photography.
Beginning graduate study with a focus on AI and machine learning while carrying forward the systems, security, and teaching perspective built during undergrad.
03 / Personal infrastructure
The lab is where curiosity becomes muscle memory.
HOME_SERVER
Windows hostDocker servicesJellyfin mediaTailscale networkRemote desktopVirtualizationSTATUS · ALWAYS EVOLVINGI maintain a Windows-based home server and keep expanding the surrounding homelab through containers, private networking, remote access, media services, and virtualization.
Alongside my Arch Linux daily driver, it gives me a place to learn by operating real systems—not only reading about them. The same patient attention shows up away from the keyboard in photography and my fascination with mechanical watches, especially Spring Drive and Seiko engineering.