ML/CV Engineer

About

We are seeking a ML/CV Engineer to support the modernization of two core production models within our computer vision platform. This is a hands-on contract role focused on model retraining and architectural improvements to significantly improve detection and classification accuracy. The engineer will collaborate closely with our internal team and operate with minimal oversight.

Project Scope & Objectives

The engineer will contribute to upgrading two existing computer vision systems:

• Vehicle Detection System
  • Current detection confidence: ~60%
  • Target detection confidence: 80%+
  • Approach may include model retraining and/or architectural redesign
• License Plate Error Classification Model
  • Current accuracy: ~90%
  • Target accuracy: 98%
  • Required high-confidence prediction rate: 80%+

The solution approach is not fully defined; the selected candidate will work collaboratively with the team to identify and implement the best technical solution.

Responsibilities

• Retrain and improve existing computer vision models to meet defined accuracy and confidence targets
• Assist with model redesign where necessary
• Build, curate, and manage training datasets
• Evaluate model performance and perform systematic tuning
• Apply best practices for model validation and accuracy improvement
• Collaborate with senior engineers while independently completing assigned tasks
• Deliver high-quality results with minimal oversight

Required Skills & Experience

• 2+ years of real-world project experience in computer vision
• Strong background in object detection and classification
• Proven experience with custom model training (not just fine-tuning)
• Solid understanding of:
  • Dataset preparation and labeling strategies
  • Model evaluation metrics and validation techniques
  • Model tuning and performance optimization
• Hands-on experience with:
  • PyTorch (required)
  • YOLO (required)
  • Vision Transformers (ViT) (required)
• Strong Python skills

Technology Stack

• Programming Language: Python
• Frameworks: PyTorch (primary), TensorFlow (optional)
• DevOps / Tooling: Azure DevOps
• Cloud Experience: AWS not required
• ML Lifecycle Tools (e.g., MLflow): Not required

Working Style & Expectations

• Ability to work with minimal oversight
• Comfortable collaborating in an exploratory environment where solutions are still being defined
• Capable of independently completing assigned tasks with high quality and reliability