Location: Las Vegas, NV
Remote: yes
Willing to relocate: depends where, prefer WFH, happy to travel on site too
Technologies: Python, PyTorch, TensorRT, Go, Linux, Docker, AWS, Claude/Codex, many more…
Résumé/CV: https://reed.pizza/cv/
Email: hn [at] reed [dot] pizza
I’m a distributed systems engineer by trade, more recently having fun getting GPUs to go brrr instead of storage drives. I am looking for full-time or contract work, in varying capacities. I typically am most interested in building out new systems, or optimizing the performance of existing ones. I prefer working in small teams, with lots of ownership over the product.
My most recent large project was to train audio localization models to detect and locate fast FPV drones (30m/s) in “realtime”. I handled dataset design and field collection, model architecture and training, and tools for testing, debugging, and visualizing all the things—the full experience. This was deployed on Nvidia jetson devices, with 3ms inference latency, with a high degree of accuracy even in new locations in a few months time.
I’m interested in applying that mix of ML and systems work beyond audio (or to audio!): getting a first model working, building training infrastructure, or making training and inference faster. Happy to own the ML side for a team new to it, or tackle a particular piece alongside people who already know the process. I’ve built training pipelines on self-hosted and cloud GPUs, and enjoy figuring out the hardware and deployment tradeoffs too.
rdallman · · focus · HN ↗
My most recent large project was to train audio localization models to detect and locate fast FPV drones (30m/s) in “realtime”. I handled dataset design and field collection, model architecture and training, and tools for testing, debugging, and visualizing all the things—the full experience. This was deployed on Nvidia jetson devices, with 3ms inference latency, with a high degree of accuracy even in new locations in a few months time.
I’m interested in applying that mix of ML and systems work beyond audio (or to audio!): getting a first model working, building training infrastructure, or making training and inference faster. Happy to own the ML side for a team new to it, or tackle a particular piece alongside people who already know the process. I’ve built training pipelines on self-hosted and cloud GPUs, and enjoy figuring out the hardware and deployment tradeoffs too.