Fesal Fayed in black and white

Hi, I'm
Fesal.

ML Engineer & a curious person.

Based in

Miami, Florida
Data Science & AI · FIU
Independent work

Portrait of Fesal Fayed

I started in premed. Now I build ML systems in Miami: small models, persistent memory, real hardware. I like work someone else can reproduce.

What interests me

The questions behind
my work.

Models under constraints

The model is only part of it. The bytes, hardware, and evaluation have to work too.

  • Video compression
  • Semantic tokens
  • Pose estimation
  • Integer entropy models
  • Tool-use tuning
  • GGUF
  • MLX
  • Quantization
  • Evaluation

Systems that remember

I want to know what happened, what failed, and what evidence a result rests on.

  • Agent workflows
  • Provenance
  • W3C PROV
  • Memory graphs
  • Negative results
  • Reproducible builds
  • Regression fixes
  • Open source

Beyond the screen

Some of the questions on my desk need hardware, not another browser tab.

  • Edge hardware
  • ESP32-S3
  • WiFi CSI
  • Computer vision
  • Person search
  • Local inference
  • GPU compute
  • Python
  • Git
  • Experiments
  • Miami
  • FIU

Featured
Projects

GitHub

Things I have released, repaired, or am still figuring out. Open a folder for the work and its status.

A non-linear path.

What I am building.
How I got here.

Right nowStill in progress

ML Engineer
& systems builder

Models, memory, hardware Miami

On my desk: provenance memory, self-healing agents, person search across long videos, and WiFi sensing on four ESP32-S3 nodes. None of these is a finished release.

Ongoing

Fixing what I use

Hermes / Topcoat / Mnemosyne

When something breaks in the tools I rely on, I want the fix upstream, not just working on my machine.

Making it usable

From training to local use

gpt-oss-20b

One tool-use fine-tune, exported for several ways of running it. The models are public on Hugging Face.

A useful constraint

How little video is enough?

comma.ai challenge Remote

The comma.ai challenge made the tradeoff concrete: fewer bytes without losing what the driving models need. The result mattered. So did making it reproducible.

The change of direction

Premed was the original plan.

Data Science & AI · FIU

I switched because this was the work I loved. It was not the straightest route, but it is the one I wanted to keep following.

Let's talk.

Working on a difficult ML problem? Building something in Miami? Or curious about an unfinished experiment? I would like to hear from you.