MB-2007Die markingrunning
Moe Babazadeh
CTO at Neuroject ↗
I architect AI-driven platforms and scalable digital products.
CTO at Neuroject, building the alignment layer for construction enterprises. Building since 2007: robots, IoT and firmware, then web, mobile and AI platforms.
I started building websites as a freelancer in 2007, studied electrical engineering with a focus on control systems, and spent my university years building robots, including a 1st place at RoboCup IranOpen 2010.
Since then I’ve built for an electronics and IoT company, a grocery marketplace, an edtech company and a construction-tech company. Today I’m CTO at Neuroject, leading the platform that connects BIM, schedules, budgets and risks for construction enterprises, and bringing AI into it.
Along the way I’ve shipped apps to the App Store and Google Play, hardware that’s still logging data after 15 years, and fixes to open source projects like Gatsby and DefinitelyTyped.
Electrical characteristics
| commits since 2021 | 11.8k+ |
|---|
| products live | 10+ |
|---|
| devices in production | 5 |
|---|
| sensor records logged | 0.5B+ |
|---|
| apps on App Store and Google Play | 2 |
|---|
| weekly downloads of packages I patched | 300k+ |
|---|
Instruction set
- TypeScript
- React & Next.js
- Laravel (PHP)
- Python & FastAPI
- AWS
- React Native & Expo
- Embedded C
- PostgreSQL & MySQL
Die spec
- Part
- MB-2007, stepping 26.10
- Die
- 11.2 × 9.6 mm · 107.5 mm²
- Stack
- FEOL + 6 metal layers + glass
- Pads
- 310, 16 bonded
The chip is a metaphor: its size and stack are made up. The numbers in this guide are real, and so are some counts on the die: 8 L2 slices for 8 skills, 8 shader cores for 8 products, 4 efficiency cores for 4 earlier roles, 50 cache mats for the 50+ Logger stations and 200 for the 200+ Attendum devices at one university.
Magnify