AI systems, embedded hardware, and computer architecture
Projects by Ahmad Saad
Case studies by Ahmad Saad, an AI engineer and Computer Engineering senior at the American University of Sharjah (AUS): two AI systems, Trend Radar and LocalAI, and two hardware and architecture builds, a Mano Basic Computer simulator and an energy harvesting LoRaWAN sensor node.
Projects
5 builds
01
Trend Radar
Production TikTok trend-detection platform where measurement decides and LLMs describe.
Impact
An estimated $60K in cost savings and a projected 20% increase in engagement rates.
- TypeScript
- Node.js
- Claude API
- PostgreSQL
- Docker
- Next.js
- REST API
02
LocalAI
Private AI for company documents that runs entirely in the browser, with page-level citations and a Playwright audit that proves nothing leaves the device.
Impact
100% retrieval and 100% single-turn answers on an integrated laptop GPU, with nothing leaving the device.
- WebGPU
- WebLLM
- Transformers.js
- TypeScript
- React
- Vite
- pdf.js
- mammoth
- IndexedDB
- Vitest
- Playwright
- GitHub Actions
- Cloudflare Workers
03
Mano Basic Computer Simulator
Cycle-accurate CPU simulator of Mano's Basic Computer with full ISA support and a CLI debugger.
- Python
04
Energy Harvesting LoRaWAN Sensor Node
Custom 2-layer PCB with an ATMEGA4809 MCU, LoRa transceiver, and an energy harvesting power path.
Impact
Sustainable, battery-free environmental monitoring powered by harvested energy.
- EasyEDA
- PCB design
- Embedded systems
05
ahmadsaad.dev
This site: a personal portfolio with a 3D coverflow project showcase, deployed on Cloudflare Workers.
- Next.js
- TypeScript
- React
- Tailwind CSS
