home
software engineer

hi, i'm mateo. i build things.
I'm a developer currently interested in AI agents, backend microservices, and cloud computing. I'm also a former SWE intern at Berkeley Lab and California State University, East Bay.
about

As a highly driven fourth-year studying Computer Science at Cal State East Bay, I am deeply passionate about the rapidly evolving world of Artificial Intelligence and Machine Learning.

My completed coursework includes: Data Structures and Algorithms, Software Engineering Operating Systems, AI, and NLP. I aspire to break into the AI sphere as a full-fledged ngineer and use my skills to develop innovative solutions to real world problems.

Software Engineer Intern June 2026 - August 2026
Lawrence Berkeley National Laboratory
Software Research Engineer Intern Jan 2026 - May 2026
California State University, East Bay
B.S. Computer Science December 2026
California State University, East Bay
A.S. Computer Science June 2024
Chabot College
skills

languages
Python · TypeScript · Java
SQL · Go · C/C++
ml & ai
LangGraph · LangChain · AI Agents
RAG · PyTorch · NLP
backend
FastAPI · REST APIs · GraphQL
Node.js · Express · Flask
database
PostgreSQL · MongoDB · NoSQL
Supabase · Neo4j· D1
frontend
React · React Native · Next.js
HTML/CSS · Tailwind · Vite
tools
Docker · AWS · GCP
Git · Linux · Slurm
projects

A multi-model coding agent integrating Perplexity web search. Engineered to use as few tokens as possible, making project generation and completion extremely cost-effective. Average project per prompt costs less than 10 cents per generated project. Winner of HackHayward 2026.
An ultra-lightweight harness TUI for coding agents using all available OpenAI models and more at only 8MB in total size. Uses an MCP server hosted via Vercel with 70+ robust tools designed to maximize agent capabilities while staying responsive. Made at the JetBrains x Codex Hackathon.
A GraphRAG research QA platform for interacting with Neo4j knowledge graphs, optimizing long-context LLM performance and reducing hallucinations. Tested and benchmarked to use up to 75.5% fewer tokens compared to other GraphRAG frameworks. Uses PolyG.
A visual speech recognition and lip-reading engine for identifying speech patterns for audibly impaired individuals. Integrates MediaPipe and OpenCV for webcam capture, video analytics, and facial overlays. Uses PyTorch and deep learning reinforcement.
A robust Java app that curates user-generated content and study sets to support personalized learning pathways with interactive review and assessments.
MYLogin is a lightweight security utility TUI written in Python and Go that allows for convenient storage and retrieval of login information with multi-layered encryption.
An automation tool for Git workflows and GitHub written in C and Bash.
A portable and compact math library written in C/C++ for working with matrices and numerical algorithms.
contact

email
aladomateo@gmail.com