Skip to content

Full-Stack & Applied-AI Engineer

AntonySaleeb

I build real-time systems and applied-AI products end to end — from orchestrated PyTorch models to the Flutter and Next.js apps that ship them.

See the work
RésuméContact

Fig. 01 — Flagship

DeepFract

DeepFract
Sample photograph being encoded
source · full bitmap
.dfc · readydetail kept
raw
enc

Input · detail retained

encode · fast
  1. Input
  2. Attention
  3. Residual
  4. Quad-tree
  5. Encoded

DeepFract is an AI-enhanced fractal image compression system. Instead of relying on a single model, it orchestrates several specialized networks that work together — residual CNNs, attention (CBAM, attention-gated U-Net), and quad-tree partitioning — so encoding stays fast while compression ratio and visual detail stay high.

Several specialised models, orchestrated — so ratio and detail both stay high without paying for it in encode time.

Classical fractal compression is slow because the encoder searches an enormous space of block self-similarities. DeepFract splits that job across specialised networks that each handle one part of the decision, then uses quad-tree partitioning to spend detail only where the image actually needs it.

Hard parts

  • 01Coordinating multiple specialized models so they improve ratio and quality without blowing up encode time.
  • 02Designing attention that surfaces structural self-similarities useful for fractal block matching.
  • 03Bridging a PyTorch inference backend with a responsive Flutter mobile client.

Stack

  • Python
  • FastAPI
  • Flutter
  • PyTorch
  • Computer Vision

Methodology & results

Evaluated on rate–distortion — compression ratio against PSNR — versus classical fractal and transform-coding baselines. The benchmark set is being re-verified against a fixed test corpus before the headline numbers go up here.

Also built

03 projects

Fig. 02 — In the app

TechTips — OS shortcuts, organised

Open case study
TechTips screen 1
TechTips screen 2
TechTips screen 3
TechTips screen 4
TechTips screen 5
TechTips screen 6
TechTips screen 7
TechTips screen 8
TechTips screen 9
TechTips screen 10
TechTips screen 11
TechTips screen 12

Fig. 03 — In the app

BT2 — numerical methods, visualised

Open case study
BT2 screen 1
BT2 screen 2
BT2 screen 3
BT2 screen 4
BT2 screen 5
BT2 screen 6
BT2 screen 7
BT2 screen 8
BT2 screen 9
BT2 screen 10
BT2 screen 11
BT2 screen 12

Background

I am a Computer Science graduate (MTI University, 2026) who builds AI-assisted, full-stack applications — from mobile UI through backend infrastructure.

I care about building things that work under real constraints, with real users and real data, not just in a tutorial. Whether it's real-time syncing for multiplayer apps, orchestrating models for a hard vision problem, or architecting a backend, I bring ideas to life with clean, maintainable code.

2026

B.Sc. Computer Science

MTI University

Software engineering, AI, and scalable full-stack applications. Active in coding competitions and hackathons.

2023 — Present

Freelance Full-Stack Developer

Self-employed

Custom web and mobile applications for clients — real-time event platforms, AI-enhanced mobile tools, and everything in between.

Antony Saleeb

Antony Saleeb

Cairo

Next.jsReactTypeScriptTailwind CSSPythonFastAPINode.jsREST APIsFlutterDartFirebasePyTorchComputer VisionAttention modelsGitGitHubSupabaseNext.jsReactTypeScriptTailwind CSSPythonFastAPINode.jsREST APIsFlutterDartFirebasePyTorchComputer VisionAttention modelsGitGitHubSupabase

Stack

Frontend

01
  • Next.js
  • React
  • TypeScript
  • Tailwind CSS

Backend

02
  • Python
  • FastAPI
  • Node.js
  • REST APIs

Mobile

03
  • Flutter
  • Dart
  • Firebase

AI / Data

04
  • PyTorch
  • Computer Vision
  • Attention models

Tools

05
  • Git
  • GitHub
  • Supabase

Contact

Open to full-stack, mobile, and applied AI roles.

Get in touch directly — no form, no gatekeeping.