Dylan Dehili

Profile

Dylan Dehili

AI engineer with an aerospace background (KNDS, Thales), I bring my ML expertise and product skills to SMEs and startups: premium websites, AI agent integration and custom automation.

Experience

Web Developer & AI Agent Integrator — Freelance

2025 — Present

Independent

Building custom websites and integrating AI agents into business workflows for SMEs and startups.

  • Modern showcase websites and web apps (Next.js, TypeScript)
  • AI agent integration to automate business processes
  • End-to-end support from requirements to production

Private Tutor & Academic Coach

Nov. 2024 — Present

Independent

Supporting middle and high school students in maths and physics-chemistry with a personalized approach.

  • Tailored sessions for each profile (confidence, conceptual gaps, exam prep)
  • Concrete examples tied to students' interests for memorable learning
  • Guidance on academic orientation and independent learning habits

Machine Learning Engineer

Sept. 2024 — Sept. 2025

KNDS France

Developing AI solutions for simulation and embedded inference in a defense context.

  • Simulation for training robots on specific tasks
  • Model quantization for edge device deployment
  • Lidar and image fusion
  • Web scraping for large-scale dataset creation

Machine Learning Engineer Intern — Computer Vision

Feb. 2024 — Sept. 2024

KNDS France

Computer vision applied to detection and segmentation, with embedded hardware deployment.

  • State-of-the-art survey on object detection and semantic / instance / panoptic segmentation
  • Pre-annotation pipeline via foundation models and Label-Studio workflow
  • Training on a 6-GPU cluster optimized for speed and accuracy
  • Real-time deployment and inference on Nvidia Xavier (edge)

AI Research Engineer Intern

May 2023 — Sept. 2023

Thales

AI research applied to electromagnetic simulation of RF components.

  • Implementation of a PINN (Physics-Informed Neural Network) for filter simulation (radial stub)
  • Transfer learning to accelerate learning across different frequencies
  • Optimization of the Residual Adaptive Refinement Mesh (RAR) method

Machine Learning Intern

Jul. 2022 — Sept. 2022

Pyxis

Developing machine learning algorithms and optimizing hyperparameters.

  • ML algorithms with TensorFlow and XGBoost
  • Hyperparameter tuning with Optuna

Education

42 Paris

Software Engineering Program

Sept. 2025 — Nov. 2026

ESTACA

Master of Engineering — Aerospace, Aeronautical & Astronautical Engineering

Sept. 2019 — Sept. 2024

Certifications

AWS

AWS Certified Machine Learning — Specialty

Amazon Web Services