I work across systems engineering, applied machine learning, and embedded software β building things where correctness and performance both have to hold under real constraints. That means reasoning carefully about concurrency, memory layout, and failure behavior, whether the target is a multi-core process, a model running entirely on-device, or a microcontroller with a few hundred kilobytes of RAM.
|
Lock-free data structures built on explicit memory ordering, cache-conscious layouts that avoid false sharing, and durable, zero-copy I/O paths β designed to stay correct under concurrent load, not just fast on paper. |
End-to-end pipelines that go from messy, real-world data to a trained model to a usable output β local-first inference, model comparison, and evaluation against ground truth, with as much attention on the engineering around the model as on the model itself. |
RTOS-based firmware with clear task separation, control logic kept independent of hardware so it can be unit-tested on a host machine, and fail-safe behavior when sensors or connectivity misbehave. |
A lock-free, low-latency message broker library in C++20, built around a cache-line-aligned SPSC ring buffer with explicit acquire/release atomics and a memory-mapped write-ahead log for zero-copy durability. CI covers sanitizers (ASan/UBSan/TSan), static analysis, and a Google Benchmark suite β performance is reported only from measured, reproducible runs.
C++20 Lock-Free SPSC mmap WAL
A local-first desktop application for monocular depth estimation. Runs PyTorch and optional ONNX Runtime models entirely on-device to generate depth maps, compare model architectures side by side, evaluate output against ground truth, and export 3D point clouds β with no cloud calls or external API keys.
Python PyTorch ONNX Runtime FastAPI Electron
An RTOS-based ESP32 room-sensing and relay-control node. FreeRTOS task separation across sensing, control, and web layers; a control-logic core kept pure and unit-tested independently of hardware; a local dashboard and HTTP API; and fail-safe relay behavior when sensors go stale or faults occur.
C/C++ FreeRTOS ESP32 PlatformIO
- Lock-free / concurrent data structures in modern C++
- RTOS fundamentals and microcontroller programming (STM32, ESP32)
- ROS 2 basics for robotics and autonomous systems
- Applied ML tooling for real-world, messy data
Open to systems, applied ML, and embedded engineering conversations β let's connect.


