Matthew Heng
Diamond Bar, California.
Hello, I'm Matthew, a junior at Diamond Bar High School. Most of my time outside class goes to research — currently at USC's Autonomous Networks Research Group, on how small a language model can get before it stops writing coherently. I've also been learning quantum software and bioinformatics, mostly by getting things wrong first. More in /research, /projects, /school, and /aboutme.
Research
-
Research Intern, Autonomous Networks Research Group (ANRG), USC · advised by Prof. Bhaskar Krishnamachari
Transformer language models from 78K to 2.9M parameters, trained on TinyStories and targeted at an ESP32. An offset power law fits the width sweep at R² = 0.998. The best sub-million model reaches lower bits per byte than the released TinyStories-1M checkpoint, though a blinded judge panel still ranks that baseline higher on story quality — compression and coherence come apart.
-
Do Large Language Models Deceive Under Pressure?
First author. Constructing simulated high-stakes conditions to test frontier models for deceptive decision making.
-
NASA GeneLab for High Schools
Twelve weeks of online space-biology and bioinformatics training run by NASA Ames' Space Biosciences Division, at roughly 20 hours a week. Omics-based analysis of spaceflight transcriptomics from NASA's Open Science Data Repository, worked through in JupyterLab notebooks. The optional Capstone Project runs by separate application, open only to students who finish the core course, and admits at most 45 teams from a cohort capped at a thousand. Qualified.
-
MIT Beaver Works, Quantum Software
Four weeks on quantum software with MITRE, admitted after prerequisite coursework in Python, systems, and the background mathematics. Built the foundation from linear algebra — complex vector spaces, unitary operators, eigendecomposition — through bra–ket notation and tensor products to superposition and multi-qubit measurement, then implemented Deutsch–Jozsa in Q# using CZ phase oracles and phase kickback. Final project: a variational quantum eigensolver for the transverse-field Ising model, written in Q# and benchmarked against exact diagonalization across ansatz depths and boundary conditions.
Projects
-
Co-founder. A nonprofit putting model rocketry kits and aerospace curriculum in front of students without access to them. I built the site that handles registration and content delivery.
-
FireShield
A Swift computer vision app that separates ordinary heat sources from genuine fire hazards, optimized for on-device inference. Presidential AI Challenge submission.
-
Nonprofit teaching Python, Java, and JavaScript to twenty-odd elementary and middle school students, through curricula built around logic and debugging rather than syntax.
School
-
Co-led biweekly meetings that introduced members to competitive programming and prepared them for hackathons including Calico at UC Berkeley; hosted BrahmaHacks, the school hackathon held each semester.
-
Helped design the curriculum for biweekly financial education meetings and coordinated participation in YIS programs including the Global Stock Pitch Competition, Finance Knowledge Bowl, Dollar-a-Day Challenge, and MyYIS Portfolio competition.
-
Researched funding opportunities and managed the club budget to support meetings, programs, and events.
-
Researched funding opportunities and managed the club budget for cultural meetings and events.
-
Managed social media outreach and recruited more than 20 new members.
-
Science Bowl
Ran team tryouts and weekly practices, using fast-paced science and mathematics questions to strengthen members' recall and problem-solving for National Science Bowl competition.
Awards
-
Gold Medalist, Calico Hackathon at UC Berkeley
Placed in the top 6% of competitors.
-
National Qualifier, FBLA Coding and Programming
Advanced from state on programming logic, algorithm design, and software development.
-
State Qualifier, FBLA Cyber Security
Qualified on network security, encryption, and digital forensics.
About me
Outside of research I play the oboe, and I'm usually the one saying yes to plans made an hour in advance. Lately I've been playing a lot of poker with friends. Mostly I like learning things I'm bad at, which is the only through-line I can find across microcontrollers, qubits, and gene expression.