Firmware Engineer @ Eaton · Boulder, Colorado 🏔️

I build and verify real-time control and communication firmware: from model, to emulator, to silicon.

while (1) { build(); verify(); ship(); }

By day I write firmware for Smart Circuit Breakers at Eaton. By night I'm usually at my bench, building things like a drum kit on an ESP32, or putting a blinky LED through four levels of verification, because "it works on my board" never felt like enough.

My path so far: vehicle motion control at ZF, emissions control software at KPIT, an M.S. at CU Boulder (with firmware internships at Danfoss, Canoo, and the SWARM-EX CubeSat), and now Eaton.

If it isn't verified, it isn't done.

• Requirements first. Tests are named after requirements and traced back to them. (see STM32H7 Control Lab)
• Verify at every level. The same code runs in a model, in an emulator, and on real hardware: MIL → SIL → PIL → HW.
• Measure, don't guess. Latency, timing, and sensor data are measured on the bench, not assumed. (see Drum Kit and Posture Tracker)

STM32H7 Control Lab: closed-loop control, verified from model to silicon I was tired of tuning gains by trial and error. This is a 7-project curriculum where OpenModelica plant models drive firmware running unmodified in Renode and on a real Nucleo-H743ZI2, with Robot Framework checking every result against written requirements. Proof: a single firmware binary passes 5/5 Robot tests on both the board and the emulator in CI, and the plant model matches the exact solution within 3.8 µV.

TI F280039C CAN Library: a portable CAN driver, built like a product CAN drivers tend to get rewritten for every project. This one has a handle-based API, a strict state machine (including ERROR_PASSIVE and BUS_OFF), hardware isolated behind a port layer, and a boundary ready for CAN FD. Proof: host-tested with GoogleTest and mocks, a GCC + Clang CI matrix, and automated releases.

ESP32 Drum Kit: press a button, hear a drum. No phone, no laptop, no Wi-Fi. I built it in six phases, each solving one constraint: interrupt-driven debounced input, a Wi-Fi AP with WebSockets, a FreeRTOS dual-core task split to remove jitter, WAV samples loaded from SD, and I2S DMA to a Class D amp. Proof: 4-voice polyphony with end-to-end latency under 10 ms.

MicroOCPP on STM32F4: the EV charger protocol on a microcontroller OCPP for EV chargers on an STM32F4, using an ESP32-C3 as a Wi-Fi bridge, with UART communication, secure WebSockets, TLS, and FreeRTOS tasks. Status: [IN PROGRESS: UPDATE WITH CURRENT STATE].

CAN Log Parser: turning raw CAN logs into answers Decodes CAN logs against a DBC file and plots any signal over a chosen time window, so you get from "something's wrong on the bus" to "here's the signal" quickly. Proof: CI, Codacy code-quality checks, and flake8 linting.

Posture Tracker: hardware that notices when you slouch Long debug sessions are hard on your back. This uses an ESP32 with an MPU-6050 to track tilt, with filtered angle output. Proof: axis maps and angle ranges verified on real hardware with recorded data, plus endurance testing.

• Adafruit Fingerprint for STM32F4: fingerprint sensor driver ported to STM32F4
• ESP32 ↔ STM32 AT Interface: AT-command link over UART, with unit and integration tests
• CANalyzer Mimic Pro: GUI for filtering and plotting CAN logs by message ID

• 🔭 Closing the first loop: the P1 thermal chamber on the STM32H7
• 🥁 Drum kit phase 5: adding an OLED display
• 🌱 Learning: Zephyr RTOS on the ESP32-S3
• 🤝 Mentoring engineers 1-on-1 on Topmate

I don’t really have an “off” switch, just a low-power mode. On weekends you’ll find me hiking around Boulder, and somewhere around mile three I start thinking about my next build.

If your CAN bus won't stop going bus-off, or you're just getting started in embedded, let's talk. I'm always happy to talk firmware over coffee ☕