Skip to content
Offshore1st

Computer Vision Engineer job description

Ready-to-use Computer Vision Engineer job description. Covers model development, MLOps, and production deployment — copy it or let us match you with pre-vetted Computer Vision engineers.

Computer Vision Engineer

1. About the role

We are seeking a skilled Computer Vision Developer to develop production-grade machine learning solutions using Object Detection (YOLO, Detectron2), Image Classification, Semantic Segmentation. This is not a research-only role — you'll take models from concept through training, validation, and deployment to production environments serving real users. The ideal candidate brings a strong foundation in statistical learning, experience with PyTorch and TensorFlow, and a track record of deploying ML systems that deliver measurable business impact. You'll collaborate with data engineers, product managers, and business stakeholders to identify high-value AI opportunities and ship solutions that scale.

2. Key responsibilities

  • Own Object Detection (YOLO, Detectron2) implementation and optimization — configuration, customization, and ongoing enhancement based on business needs
  • Manage Image Classification workflows including setup, user training, and continuous improvement of processes
  • Implement and maintain Semantic Segmentation ensuring seamless integration with existing systems and workflows
  • Design and train machine learning models using Computer Vision best practices and modern architectures
  • Build and maintain data pipelines that feed ML models with clean, validated training data
  • Deploy models to production with monitoring, alerting, and automated retraining capabilities
  • Collaborate with product teams to identify high-impact AI use cases and estimate feasibility
  • Conduct model performance analysis — precision, recall, latency, and business impact metrics
  • Document model architectures, training procedures, and serving infrastructure for team knowledge sharing
  • Stay current with Computer Vision advances and evaluate new techniques for potential adoption

3. Must-have qualifications

  • Hands-on experience with Object Detection (YOLO, Detectron2) — configuration, customization, and troubleshooting in production environments
  • Proficiency with PyTorch as part of the Computer Vision development/operations workflow
  • 3+ years of Computer Vision experience with models deployed to production environments
  • Strong foundation in statistics, linear algebra, and probability theory
  • Experience with the full ML lifecycle — data preparation, training, validation, deployment, and monitoring
  • Proficiency in Python and common ML frameworks (TensorFlow, PyTorch, or scikit-learn)
  • Understanding of MLOps practices — versioning, reproducibility, and automated pipelines

4. Nice-to-have skills

  • TensorFlow Developer Certificate certification or equivalent validated credential
  • AWS Machine Learning Specialty certification or equivalent validated credential
  • Experience with advanced Computer Vision features: Image Classification, Semantic Segmentation, OCR & Document Processing
  • Familiarity with the broader Computer Vision ecosystem including TensorFlow and OpenCV
  • Published research, conference presentations, or open-source ML contributions
  • Experience with large language models (LLMs), RAG architectures, or generative AI applications

5. Interview tips

End-to-End ML Design

Present a business problem and ask the candidate to design a complete Computer Vision solution — data requirements, model approach, evaluation metrics, and deployment plan.

Code Review Exercise

Show them a Computer Vision code snippet with subtle issues (data leakage, incorrect validation split). See if they catch them and explain the implications.

Production Debugging

Describe a scenario where a deployed Computer Vision model's performance has degraded. Ask them to walk through their diagnostic process step by step.

Technical Presentation

Ask them to present a past Computer Vision project — problem, approach, results, and what they'd do differently. Evaluate technical depth and communication clarity.

6. Typical team structure

Team size
2-4 Computer Vision engineers
Reports to
Head of AI/ML, VP of Engineering, or CTO
Collaborates with
Data Engineering, Product Management, Backend Engineering, Business Intelligence

Related job description templates

All job description templates

Skip the JD. Get matched instead.

Pre-vetted Computer Vision profiles with video intros in 24 to 48 hours.

You'll get 3 to 5 profiles within 48 hours. Nothing is signed until you've interviewed.