Hey, I'm John!
I am a current Junior studying Computer Science at UC Irvine set to graduate with a Bachelor's by June 2028. I'm looking for software engineering internships and new grad roles, with a focus on full stack, machine learning, and generative AI.
Currently I'm a Software Engineer Intern at TrellisWare Technologies, where I work on the data and infrastructure behind their radio hardware testing — building pipelines and tooling that make test results trustworthy and fast to query. Before that I was at My Connective Care, building a full stack platform for patient care coordination and the deployment pipeline behind it.
In my free time, I like to surf, go on hikes and play the guitar. I love working with other people and making new connections — if any of this sounds like your team, my résumé and links are below, or just email me at john.martian2024@gmail.com and let's talk.
- Engineered a Python/PostgreSQL ETL pipeline unifying 1000+ JSON/CSV/.log automated test results for radio hardware verification, cutting manual validation by 80%
- Implemented nginx as a TCP reverse proxy in front of PostgreSQL, routing all client connections through a single entry point and eliminating direct external access
- Built database push into the existing Ansible/Fabric/pytest automation as a custom pytest CLI option, enabling one-command ingestion and cutting query times by 50%
- Built a full-stack site improving patient-facing care coordination in React, TypeScript, and Fastify, consolidating caregiver, clinician, and family updates into one role-based platform
- Built CI/CD pipelines with GitHub Actions and Docker, deploying to AWS (Cognito, Amplify, EC2) at 100% uptime
- Supported 200+ students by troubleshooting C/C++ code, clarifying core concepts, and answering real-time questions
- Created supplemental reference materials and walkthroughs for complex data structure implementations
- Built ML pipelines in PyTorch for amino acid sequence detection, training Spiking Neural Network architectures on biological datasets
- Developed data visualization and analysis tooling with NumPy and Matplotlib to evaluate model performance across perception and attention tasks
Selected work
04 PROJECTS · CLICK TO OPENAn audio-to-tablature pipeline: Spotify's Basic Pitch neural network transcribes a recording into notes, then a custom dynamic-programming pass picks the fret positions a human hand would actually choose — minimizing travel across the neck instead of just matching pitch.
A personal platform for writing album reviews, styled after the listening experience itself: cards flip to reveal the writeup, Postgres holds the catalog, and Resend fires email notifications to subscribers whenever a new review lands.
Fine-tuned OpenAI's VPT model to play Minecraft toward a single long-horizon goal: mine, smelt, and hold an iron ingot without human input. Reward shaping and behavioral cloning on top of the pretrained policy carried it to a 95.6% success rate.
Three competition websites for a synthetic-biology team, one per season, each carrying a year of wet-lab and modeling research from multiple departments into something judges could actually navigate. Gold, Silver, and Bronze medals against a field of 100+ schools.
Let's paddle out.
Open to software engineering roles and internships. Fastest way to reach me is email.