David Brin

The short version

Computer engineer, currently building an AI company. The pattern started early and hasn't changed much: take the thing apart, find out how it actually works, then build one. That has meant robots, a garage band, PCBs for aircraft that had to fly, neural recordings, embedded C that ships to hardware, and now a platform that makes a company's own knowledge useful to a model.

A sentence carved into a school desk still runs underneath all of it: don't find your place in the world, make it. The long version of that story is The Path, a river you scroll from the first robotics kit to now.

Now

Education

Previously

• General Atomics (Jun 2025 – Jun 2026): electrical technologies intern — embedded firmware in C for a camera driver and control system, shipped through Git/Bitbucket feature branches and pull requests with Linux tooling and hardware validation; optical-controls experiments alongside optical scientists, and development boards configured for system testing and integration.
• Voytek Lab, UC San Diego (Apr 2024 – Jun 2025): undergraduate researcher building neural-data pipelines for patch-clamp action potentials and organoid multi-electrode-array field potentials, in MATLAB scripts and Python notebooks.
• Triton Unmanned Aerial Systems (Sep 2023 – Apr 2024): embedded-systems project team — PCB design in Altium Designer and mechanical / electrical design in AutoCAD for aircraft that had to actually fly.
• Berkeley Coding Academy (Jul – Aug 2022): cohort lead, teaching intro data science and machine learning in Python and Google Colab notebooks.

Selected projects

• Cross-teaching segmentation (Fall 2025) — a U-Net and a ViT grading each other's pseudo-labels on 22 labeled micro-CT slices. The demo runs the real checkpoints and curves.
• Nocturnal Neuro (Winter 2024) — an EEG-based ML diagnostic-device concept for bipolar disorder, spanning hardware, signal processing and the technical pitch: reworked board, schematic and BOM, and a real recording with the DSP pipeline live.
• IEEE HardHacks 2026 (Jan 2026) — a microcontroller-based embedded security system talking to a mobile app over MQTT, built in a weekend and rebuilt here as one connected simulation: house, firmware, MQTT, phone.
• Data Refinement Pipeline (Jan 2026) — a web-based refinery platform for organizing and improving data assets, built in a sprint in the run-up to founding Katalyxt.
• Organoids on psychedelics and Anatomy of a spike — the Voytek Lab work, replayed on real public recordings.
• Autonomous car (Spring 2025) — ROS 2 on Linux with onboard NVIDIA compute and a Roboflow vision stack. No interactive rebuild yet; it sits on The Path as an honest gap.

18 demos are playable in the browser — see all the demos, or read the write-ups on David's Wikipedia.

Skills

• Python
• C / C++
• Java
• TypeScript
• MATLAB
• Verilog / VHDL
• LangChain
• LLM agents & orchestration
• Context / memory systems
• n8n
• PyTorch
• Deep learning
• Computer vision
• Embedded systems
• AWS
• Azure
• Linux
• Git / Bitbucket
• IAM / Auth0
• Claude, Codex & Cursor
• ROS 2
• MQTT
• Altium

Elsewhere

• GitHub — where the research code lives.
• katalyxt.ai — the company.
• The Path — the same story as a river you scroll.
• The demo video — 23 seconds of this site, hosted on the YouTube replica.
• david.e.brin@gmail.com — the fastest way to reach him.