Modern AI systems face a critical energy constraint, so they either rely on cloud processing (creating latency and privacy concerns) or drain batteries quickly with on-device
Open presentation image | Open presentation video
Modern AI systems face a critical energy constraint, so they either rely on cloud processing (creating latency and privacy concerns) or drain batteries quickly with on-device computation.
During the scope of Fab Academy, I'm exploring the fundementals of neuromorphic computing systems that could develop into a later, more robust solution for that issue. My target is to design and fabricatecustom brain-inspired circuit boards that use event-based processing to visualize how neurons work and use brain-inspired software processing within different applicaions.
Right now, I'm exploring different applications for these neuromorphic/neuromorphic-like circuits from robotics to AR/XR systems to brain-computer interfaces, while learning about their capabilities and constraints through hands-on development at Fab Academy.
Throughout the program, I'll be developing, refining, and documenting the complete journey from initial concept to final fabrication.
I'm building the project in three spirals. Spiral 1 is a baseline wearable due by the end of system integration week (May 6), which is a self-deadline xD. Spirals 2 and 3 layer on top of it after that.
Forehead enclosure with the electrodes and ATtiny, XIAO ESP32-S3 on the frame, button plus 5 capacitive pads, camera, speaker, color display with optics, external 5000 mAh powerbank. Software does standard EEG processing (filter, band power, blink detection) and renders it on the display. XIAO talks directly over WiFi to my personal-assistant endpoint, which brokers Gemini and my other agents and shares memory.
Layered onto spiral 1 hardware, same physical build:
Full physical rebuild, software from spiral 2 stays:
Beyond the final-project scope: separate EMG/ECG hand controller for driving the display (future work).
| Week | Dates | Topic | Progress and tasks for final project |
|---|---|---|---|
| W02 | Jan 21-26 | CAD |
|
| W03 | Jan 28-Feb 4 | Computer-controlled cutting | ✓ Multi-layer logo sticker |
| W04 | Feb 9-11 | Embedded programming |
|
| W05 | Feb 18-23 | 3D scanning and printing |
|
| W06 | Feb 25-Mar 4 | Electronics design | ✓ LIF neuron PCB in KiCad |
| W07 | Mar 9-11 | Computer-controlled machining | ✓ Non-project build, skills week |
| W08 | Mar 18-23 | Electronics production | ✓ Milled and stuffed my first ever PCB |
| W09 | Mar 25-Apr 1 | Input devices |
|
| W10 | Apr 8-13 | Output devices |
|
| W11 | Apr 15-21 | Networking and communication + midterm review |
|
| W12 + break | Apr 22-27 | Mechanical / machine design |
|
| W14 | Apr 29-May 6 | Molding and casting, spiral 1 work |
|
| W15 | Apr 29-May 6 | Interface programming + first DIY EEG test bench |
|
| W16 | May 13 | System integration, physical spiral 1 |
|
| W17 | May 20-25 | SNN dashboard and display spiral |
|
| W18 | May 27 | Applications, implications, and final scope |
|
| Final | Jun 8-12 | Project presentations |
|
I set up my working environment, learned the Git/GitLab workflow, rebuilt the first version of my personal site, and opened the first notes for the final project.
I used CAD week to design the first visual identity and glasses concepts for my final project, moving between vector work, 3D modeling, and file compression.
I made vinyl stickers, cut laser parts, and turned the neuron-model idea into a press-fit kit while learning the cutting settings that actually matter.
I explored embedded programming through XIAO ESP32-S3 tests: stepper and blink experiments, Grove Shield soldering, sensor input, OLED animation, camera streaming, and Muse EEG experiments.
I printed and scanned objects while pushing toward the final project: a Mobius lattice test, glasses hangers for the XIAO Grove Shield, and a Hyperscape scan of the lab.
I designed my first PCB around a hardware LIF neuron idea, moving from circuit explanation to KiCad schematic, board layout, and manufacturing checks.
I learned CNC machining by designing Syrian mosaic geometry and a Damascus Gate-inspired panel, then moving the model through Fusion CAM and machining prep.
I simplified and produced a PCB, learned how in-house milling rules affect board design, and compared that path with board-house production for future project boards.
I focused on input devices by redesigning the neuron PCB as a software LIF board, then milling, soldering, and testing it as an input-focused electronics step.
I built output-device experiments around the glasses system, including a 3D PCB concept with display, buzzer, and LED, then worked through milling, soldering, and display issues.
I documented networking through ESP-NOW group work and built a networked music-player direction using the XIAO, a receiver board, and the broken-OLED lessons.
I contributed to the machine build by editing CAD, testing PLA wood filament, and working on the frontend/backend link for the drawing-machine interface.
I used the midterm review to organize the systems diagram, final-project plan, and remaining schedule before the final spiral of work.
I explored molding and casting through wax-stamp, fridge-magnet, low-melt metal, and biomaterial tests, then compared the processes and what each one is good for.
I built the software side around XR machine interfaces and the NeuroAR dashboard, including VRKanji, ARBrushMachine, and the first agentic glasses interface ideas.
I integrated the first NeuroAR spiral by building the chassis, optics, electronics, milled board, and final packaged glasses prototype.
I used wildcard week for embroidery, heat pressing, and sandblasting: a logo test, a Syria design, and a few fabrication processes outside the normal electronics/CAD path.
I clarified the applications and implications of NeuroAR: what it will do, who it is for, what I need to build, what it costs, and how I will evaluate it.
I wrapped Fab Academy by documenting invention, IP, income, and how I plan to release NeuroAR as an open project with the FAB license and future spirals.
