$ whoami
ECE @ UT Austin · AWS Software Engineer Intern · building edge-AI, embedded & ML.
I build at the boundary of software and hardware — from STM32 firmware to agentic AI developer tools.
// about
Hello!!! I'm Kaitlyn, an ECE student at UT Austin. I work in the overlap between embedded systems and machine learning — close enough to the hardware to care about clock cycles, close enough to the models to care about what they get wrong. Lately that's looked like a machine-learning steering controller for an autonomous car, computer vision that runs offline on modest hardware, firmware for a hyperloop pod, a stint as a software engineering intern at Amazon building agentic AI tooling, and, for no defensible reason, a working Fruit Ninja game on a microcontroller — running on a custom PCB I designed, printed, and soldered myself!
I love watching code I wrote reach out and move something in the real world, and I'm just as curious about pushing AI down onto that same hardware, where it has to be small, fast, and right the first time. I tend to chase ideas all the way to something that actually works: an offline translation tool I built ended up with real users requesting features, which is still the most fun kind of feedback I've gotten. Being a first-gen student taught me to reverse-engineer systems nobody explained to me, which is conveniently most of engineering. I'm just as comfortable training a model in Python as I am a few layers below where most software stops — in the firmware, in the signal, in the part where it either works on the bench or it doesn't.
Off the clock, catch me at more concerts than is reasonable, playing sudoku and learning more patterns to solve them faster, hiking, or working on my steady rotation of Raspberry Pi and Arduino projects :D!
// experience
TypeScript · Java · MCP · AWS Lambda · CloudWatch
C/C++ · MATLAB/Simulink · Python · Gaussian Process Regression
C · STM32 · ADC · DMA · UART · CAN
// projects
Open-source tool that diagnoses why ML models break or slow down when deployed to edge hardware.
Profiles a model against a target device and surfaces the real bottlenecks — quantization mismatches, unsupported ops, memory pressure — instead of leaving you to guess.
Data-driven steering controller for an autonomous vehicle, built on Gaussian Process regression.
Trained a GP model on a 678k-point system-identification pipeline to control steering, and used the GP's own variance estimate as a live fault detector — flagging when the model was operating outside its trusted region.
→ 0.23 m RMSE tracking error.
Offline facial-recognition / computer-vision tool.
Real-time multi-face identification using 128-dimensional ResNet embeddings — runs fully offline with no cloud dependency.
Offline speech-to-text + translation pipeline.
End-to-end transcription and translation that runs locally. Open-sourced and picked up by real external users.
Full-stack LLM app serving generated content with safety guards.
FastAPI + PostgreSQL backend serving LLM-generated content, with structured-output validation and fallback guards so a bad model response never breaks the experience.
// skills
// contact
I'm looking for software / ML / embedded / edge-AI internships. The fastest way to reach me is email.