Skip to content
Currently at Arago — open to chats on AI compilers, accelerators, and robotics.

Alhussein Jamil Software Engineer — AI Systems

Software engineer at Arago. I work on the stack that runs PyTorch on a custom optical AI accelerator.

See my work Download CV Get in touch →
Portrait of Alhussein Jamil, a software engineer with short dark hair, wearing a navy suit, white shirt, and patterned blue tie, photographed against a neutral grey background.
01 · About

About

Software engineer at Arago, working on the stack that runs PyTorch on a custom optical AI accelerator — compiler passes, runtime, kernels, and the ISA that ties them together.

Earlier: reinforcement-learning control policies for bipedal robots and exoskeletons at Wandercraft, and computer-vision pipelines for industrial automation at EyePick. École Polytechnique, 2024.

02 · Experience

Experience

  1. Software Engineer — AI Accelerator Stack · Arago

    Paris, France

    Oct. 2024 — Present

    • Develop core software components for a custom optical AI accelerator focused on matrix multiplication.
    • Build a PyTorch-to-hardware execution path: graph lowering, tensor transformation, command generation, and runtime orchestration.
    • Implement low-level kernels and ISA-level execution flows for a proprietary multi-core accelerator architecture.
    • Integrate compiler, runtime, driver, and hardware-facing components into end-to-end ML accelerator demos.

  2. Machine Learning Engineer · EyePick

    Paris, France

    Sept. 2023 — Sept. 2024

    • Built computer-vision pipelines for real-time robotic automation in industrial, agricultural, and culinary environments.
    • Built image-based anomaly detection and classification systems for quality-control applications.
    • Adapted ResNet-based models as an alternative to YOLO-based detection pipelines under licensing constraints.

  3. Reinforcement Learning Intern · Wandercraft

    Paris, France

    Mar. 2023 — Sept. 2023

    • Trained reinforcement-learning control policies for the Cassie bipedal robot and the Eve exoskeleton.
    • Used NVIDIA Isaac Gym and Ray RLlib for parallel simulation and distributed policy training.
    • Worked on sim-to-real transfer using imitation learning, domain adaptation, and visual adaptation methods.

  4. AI Algorithms Intern · Poppins (formerly Mila)

    Paris, France

    Jun. 2022 — Sept. 2022

    • Implemented an adaptive-difficulty algorithm for a therapeutic game designed for dyslexic children.
    • Used clustering and gameplay metrics to adjust difficulty dynamically without explicit child feedback.
    • Contributed to algorithmic development and game-side integration.
03 · Projects

Projects & Awards

Cassie Bipedal Control

Reinforcement-learning environment and control policy for the Cassie bipedal robot, built on Ray RLlib and trained in parallel simulation.

  • RL
  • Robotics
  • Ray RLlib
View on GitHub

ViT with Hilbert-Curve Patches

Vision Transformer variant that orders patches along a Hilbert space-filling curve to better preserve 2D locality in the token sequence.

  • Vision
  • Transformers
  • Research
View on GitHub

Convolutional Wasserstein Distances

Implementation of fast optimal transport on geometric domains via heat-kernel convolutions, after Solomon et al.

  • Optimal Transport
  • Geometry
  • Python
View on GitHub

Bahdanau Attention NMT

From-scratch PyTorch implementation of the Bahdanau attention mechanism for neural machine translation.

  • NLP
  • Attention
  • PyTorch
View on GitHub

U-Net Source Separation

U-Net trained to separate audio sources from mixed spectrograms.

  • Audio
  • Deep Learning
  • PyTorch
View on GitHub

Haptic Rendering — Final-Year Project

Real-time haptic rendering pipeline in C++ and GLSL, developed as an École Polytechnique research project.

  • Haptics
  • C++
  • GLSL
View on GitHub

Scientific Olympiads

Three-time Silver Medalist at the Syrian National Olympiads in Mathematics, Informatics, and Robotics.

  • Award
  • Mathematics
  • Informatics
  • Robotics
04 · Gallery

3D Creations

A few personal 3D renders.

05 · Skills

Skills

Programming

  • Python
  • C++
  • C
  • OCaml
  • Java
  • C#
  • SQL

AI Systems

  • PyTorch execution
  • Graph lowering
  • Runtime design
  • Custom ISA
  • Kernels
  • Multi-core execution

Machine Learning

  • Computer vision
  • Reinforcement learning
  • Imitation learning
  • Sim-to-real
  • Model deployment

Robotics

  • NVIDIA Isaac Gym
  • Ray RLlib
  • Robotic control
  • Industrial automation
  • VR robot interfaces

Tools

  • Linux
  • Git
  • Docker
  • Debugging
  • HW/SW integration
06 · Education

Education

2020 — 2024

École Polytechnique

Engineering Degree — MSc equivalent

  • · Mathematics, physics, statistics, machine learning, and robotics.

2023 — 2024

Sorbonne University — UPMC

Master 2 — Intelligent Systems

  • · Built a VR interface for controlling and visualising a Franka Emika robotic arm.

2018 — 2020

Lycées Louis-le-Grand & Buffon

CPGE — MPSI → MP*

  • · Intensive preparatory classes in mathematics and physics.
08 · More

Languages · Teaching · Interests

Languages

  • Arabic Native
  • French Bilingual
  • English Bilingual
  • Spanish Intermediate

Teaching

  • X-HEC Master

    Guest lecturer in machine learning for engineering and business master's students.

  • Lycée Saint-Louis

    Oral examiner in mathematics for CPGE students.

Interests

  • MMA
  • Puzzles
  • Violin
  • Climbing
09 · Contact

Contact

Say hi

The form is routed through Cloudflare. You can also click the address directly — I usually reply within a few days.

LinkedIn GitHub

Send a message

Routed through a Cloudflare Pages Function — no mail client opens.