ML Technical Lead – Computer Vision & Shelf Recognition

Décryptage du poste par Postule AI

Résumé du rôle

Zsystem is hiring a hands-on ML Technical Lead to spearhead the development of a next-generation shelf recognition system. The role involves leading a CV/ML engineering team, designing YOLO-based training pipelines, and deploying models into retail applications. Candidates need 5+ years in computer vision with 2+ years in leadership, strong PyTorch and AWS expertise, and experience with synthetic data and object detection. This is a technical leadership position combining deep ML engineering with team mentorship and cross-functional collaboration.

Exigences clés (estimation)

Lead a small CV/ML engineering team and build the roadmap for detection and recognition models.
Design and optimize training pipelines for YOLO-based models using real and synthetic datasets.
Implement best practices for data collection, augmentation, annotation quality, and tiling for small objects.
Establish CI/CD workflows for model training, versioning, deployment, and A/B testing.
Mentor junior engineers and promote a strong ML engineering culture.
Collaborate with product and operations teams to deploy models into retailer-facing applications.

Compétences

5+ years in computer vision/deep learning, including 2+ years in a lead role.Strong experience with PyTorch, Ultralytics YOLO, and distributed training on AWS.Expertise in synthetic data generation and annotation pipelines.Experience with object detection on small objects, data imbalance, and augmentation strategies.Excellent communication and cross-functional leadership skills.Ability to explore and evaluate advanced approaches like OBB vs. HBB and DINOv3 backbones.

Niveau estimé

This role requires 5+ years of experience in computer vision and deep learning, with at least 2 years in a technical leadership position. Candidates should have proven expertise in leading ML projects and teams.

Attributs détectés par l’IA (si absents de l’offre)

Type de poste: CDI
Niveau d'expérience: 5–10 ans
Fonction: Informatique / Ingénierie
Secteur: Informatique / IT
Langues: Anglais

Cette analyse a été générée automatiquement par Postule AI à partir de l'offre.

Introduction to the Position

We're seeking a hands-on ML Technical Lead to lead the development of our next-generation shelf recognition system using YOLOv8, synthetic data workflows, and potentially DINOv3-based architectures.

Your Role

Lead a small CV/ML engineering team and build the roadmap for detection and recognition models.
Design and optimize training pipelines for YOLO-based models (real + synthetic datasets).
Implement best practices for data collection, augmentation, annotation quality, and tiling for small objects.
Explore and evaluate approaches such as OBB vs. HBB, DINOv3 backbones, and multi-GPU distributed training.
Establish CI/CD workflows for model training, versioning, deployment, and A/B testing.
Mentor junior engineers and promote strong ML engineering culture.
Collaborate with product and operations teams to deploy models into retailer-facing applications.

Your Qualifications

5+ years in computer vision/deep learning, including 2+ years in a lead role.
Strong experience with PyTorch, Ultralytics YOLO, and distributed training on AWS.
Expertise in synthetic data generation and annotation pipelines.
Experience with object detection on small objects, data imbalance, and augmentation strategies.
Excellent communication and cross-functional leadership.

Bonus

Experience with DINO/DINOv2/v3, ViT-based backbones.
Background in retail tech, OCR, or dense shelf detection.
On-device optimization experience (TensorRT, TFLite, etc.).

Recruitment Process

CV pre-screening
AI Interview
Technical and fit interview with Zsystem team

Cette description d'emploi a pu être reformatée par Postule pour améliorer sa lisibilité et sa présentation. Le contenu et les informations restent fidèles à l'offre d'emploi originale. .