Summary
Overview
Work History
Education
Skills
Timeline
Generic

Dylan Dehili

Paris

Summary

Offering solid foundation in scientific principles and strong desire to learn and develop in research environment. Brings keen ability to understand complex problems and implement technical solutions such as data analysis and computational modeling. Ready to use and develop analytical and problem-solving skills in machine learning role.

Overview

3
3
years of professional experience

Work History

AI Research engineer -GenAI/Computer Vision

Dassault systems
10.2024 - Current
  • Utilization of Wonder3D & Gaussian Anything to transform textual descriptions or images into preliminary 3D models using natural language processing (NLP) algorithms and generative networks, accelerating the initial mechanical parts design phase by 80%
  • Python, TensorFlow/PyTorch, and CAD libraries (e.g., OpenSCAD, FreeCAD API) to develop design automation scripts

AI Engineer intern - NLP/Computer Vision

KNDS France (Nexter robotics)
02.2024 - 08.2024
  • Conduct a state-of-the-art survey in Computer vision focusing on object detection, semantics segmentation, instance segmentation and panoptic segmentation
  • Developed a Labelling pipeline using foundation models and managed the annotation workflow with Labelstudio and kili
  • Train selected model on a cluster of 6 GPU with the newly annotated data
  • Deployed and ran inference on edge devices (Nvidia Xavier)
  • Create a RAG application using Langchain and Gradio to build a technical documentation helper

AI Research Engineer intern

Thales Research and Technology
05.2023 - 10.2023
  • Implementation of a PINN (physics-informed neural networks) for electromagnetic simulation of a transmission line filter (radial stub)
  • Employed transfer learning to accelerate model learning over different frequencies, facilitating faster, more efficient simulation of RF components across a wide range of operating conditions
  • Implemented and optimized the Residual Adaptive Refinement Mesh (RAR) method, dynamically adapting the simulation grid to focus on regions where prediction errors were highest

Machine Learning Engineer intern

Pyxis
07.2022 - 09.2022
  • Creation of machine learning algorithms to classify web traffic with Tensorflow and Xgboost
  • 97% accuracy and 96% recall sur 8 classes
  • Use Optuna for hyperparameter tuning

Education

Master of Science - Aerospace Engineering

ESTACA
Paris
09-2024

Skills

  • Effective multitasking
  • Team collaboration
  • Written/Verbal communication
  • Idea generation
  • Python programming

Timeline

AI Research engineer -GenAI/Computer Vision

Dassault systems
10.2024 - Current

AI Engineer intern - NLP/Computer Vision

KNDS France (Nexter robotics)
02.2024 - 08.2024

AI Research Engineer intern

Thales Research and Technology
05.2023 - 10.2023

Machine Learning Engineer intern

Pyxis
07.2022 - 09.2022

Master of Science - Aerospace Engineering

ESTACA
Dylan Dehili