You're seeing this page as if you were . The main menu is still yours, though. Exit from immersion
Louis MartinezLM

Average response time: 1 hour

Freelancer profile translated to English.
Back to original language

About Louis

A graduate of Télécom Paris and ENS Paris-Saclay, I operate at the intersection of demanding academic research and the strict constraints of software engineering.
My goal: audit your AI infrastructures, rework your Computer Vision pipelines, and deploy your models to make them robust, fast, and scalable solutions.

What I concretely bring to your projects:

  • Audit & Redesign of CV Architectures: Analysis of your models, identification of bottlenecks (latency, costs), and rework for clean, maintainable pipelines.
  • Deployment under constraints (Real-Time & Edge): Inference optimization to ensure low latency on Cloud (RunPod, HPC) or on local hardware.
  • Iterative Execution: Rapid prototyping and delivery of "plug & play" solutions.

Expertise proven by results:

  • Production Reliability: Optimization and deployment of AI models for the LivePortal real-time video installation, used during events for Microsoft, PSG, and Paris Creator Week.
  • Software Performance: 3x reduction in hardware testing time in factories through the creation of automation tools (Sagemcom).
  • Edge Computing (Open Source): Google Summer of Code contributor for the development of a highly optimized real-time CV pipeline (inference on laptops).

  • Scientific Validation: Design of complex Deep Learning architectures (3D Transformers) published at CVPR 2026, the #1 global conference in Computer Vision.
  • Execution Under Pressure: Multiple podiums at international AI hackathons (Doctolib, Hugging Face, Anthropic).

Stack:

Python, PyTorch, Hugging Face, Docker, Bash, Git, Cloud GPU.
  • French

    Native or bilingual

  • English

    Fluent

Remote only
Primarily works remotely

Experience

  • Obvious Research
    ML & Computer Vision Engineer
    March 2026 - Today (5 months)
    Puteaux, France
    Development, optimization, and deployment of LivePortal, a real-time generative video AI art installation interacting via voice and text. Mission focused on critical performance, cloud/local deployment, and continuous model improvement for highly demanding international events.

    Key Achievements

    • Real-Time Deployment & Performance: Production deployment (Docker) and inference optimization of video generation models to ensure absolute fluidity (constant 30FPS) and minimal latency in live conditions.
    • Robustness Engineering: Securing the software infrastructure through containerization to transform a research model into an ultra-reliable "plug-and-play" product.
    • Model Improvement & Adaptation: Reworking the AI architecture (based on the Hugging Face ecosystem) to continuously enhance the visual quality of the generation.
    • GPU Infrastructure Management: Provisioning and orchestration of computing instances (Cloud GPUs, HPC clusters) for training and inference.
    • Reliability Proof: Successful deployment at world-class events for leading clients: Microsoft, Paris-Saint Germain (Champions League semi-final), and Paris Creator Week.
    Hugging Face RunPod (Cloud GPU) Docker HPC PyTorch
  • LIX, École Polytechnique
    3D Computer Vision & Deep Learning Research Engineer
    April 2025 - February 2026 (10 months)
    Palaiseau, France
    End-to-end design and development of an Artificial Intelligence solution for spatial data analysis (3D point clouds). This project required complete mastery of the Machine Learning lifecycle, from data creation to scientific validation.

    Key Achievements

    • End-to-End Management: Full management of the AI pipeline, from data acquisition and structuring to final model design.
    • Data Engineering & Evaluation: Creation of a custom benchmark to rigorously test and audit models (key skill for ensuring production reliability).
    • Advanced Modeling & Efficiency: Development of Deep Learning architectures (Transformers). Strategic use of pre-trained models to maximize prediction performance despite limited data volume.
    • Engineering & Containerization: Setting up environments using Docker and optimizing training on local compute clusters.

    • Scientific Excellence: Work validated and accepted at CVPR 2026, the #1 global conference in Computer Vision.
    PyTorch Python Docker Computer Vision Deep Learning
  • Université McGill
    ML & Computer Vision Engineer (Google Summer of Code)
    March 2024 - September 2024 (6 months)
    Montreal, QC, Canada
    Contributor to the Google Summer of Code program. End-to-end development of a Computer Vision pipeline translating human movements into music in real-time. The main challenge was architectural optimization to ensure low-latency local execution.

    Key Achievements

    • Data Lifecycle Management: Full project management, from raw data collection strategy to model training and evaluation.
    • Optimization & Edge Computing: Reworking AI architectures and developing automation scripts (Bash) to ensure real-time, very low-latency inference directly on standard laptops.
    • Product & Client Orientation: Close collaboration with end-users (musicians and contemporary choreographers) to translate their business requirements into technical specifications and adapt the model's behavior.
    • Software Excellence: Delivery of a turnkey, robust, and fully open-sourced solution, marking the official success of the mission for the Google program.
    Real Time Python PyTorch Bash Computer Vision

Recommendations

Be the first to recommend Louis

Help this freelancer shine by sharing your experience working together.

These freelancer profiles also match your criteria

AgathaA

Agatha Frydrych

Backend Java Software Engineer

4.7

(3)

2

BaptisteB

Baptiste Duhen

Fullstack developer

4.6

(4)

5

AmedA

Amed Hamou

Senior Lead Developer

4

(2)

7

AudreyA

Audrey Champion

Web developer

4.3

(3)

4

Education

  • Master 2, MVA (Mathematics, Vision, Learning)
    École Normale Supérieure Paris-Saclay
    2025
    Master 2 en maths appliquées au Machine Learning et à la vision par ordinateur, Mention Très Bien
  • Engineering Degree
    Télécom Paris
    2025
    Cursus ingénieur en Traitement du Signal & Vision par Ordinateur, Mention Très Bien

Skill set

Categories