Senior AI / Computer Vision Engineer

About the Client

Our client is North America’s leading network of independent aftermarket truck parts distributors. The customer`s distributors serve the needs of their clients from more than 700 locations across the United States, Canada, Puerto Rico, and Mexico. The customer`s distributors are specialists who understand the demands of their local, regional, and national clients for quality parts and exceptional service.

Our client is a proud member of NEXUS North America and NEXUS Automotive International, a worldwide group of parts distributors committed to bringing a global approach to the automotive and commercial vehicle aftermarket industries.

About the Role

We are building a visual recognition solution for automotive parts that enables the identification of parts using mobile phone images. This is a complex, real-world problem involving large-scale catalogs, visually similar items, and non-ideal input conditions (dirty, worn, or partially visible parts).

We are looking for a Senior AI / Computer Vision Engineer who will take ownership of the AI component of the solution — from model design and experimentation to production-ready implementation and continuous improvement.

This role is highly R&D-driven, requiring strong technical expertise combined with a pragmatic mindset and the ability to adapt approaches as new challenges arise.

What Success Looks Like

• Delivers a working visual recognition approach for selected part categories
• Continuously improves model accuracy using real-world data
• Effectively handles edge cases and visually similar parts
• Adapts strategy when initial assumptions do not hold
• Contributes to building a scalable and maintainable AI pipeline

Why This Role Is Interesting

• Work on a high-impact, real-world problem with direct business value
• Solve challenges at the intersection of AI, data, and operations
• Be part of building a system from early-stage validation to a scalable solution
• High level of ownership and influence on technical direction

Requirements:

• 5+ years of experience in Computer Vision / Machine Learning
• Strong hands-on experience with PyTorch, TensorFlow, or similar frameworks
• Experience with image-based deep learning models (CNNs, transformers, embeddings)
• Experience with model training pipelines and evaluation methodologies
• Solid understanding of data quality, dataset design, and augmentation techniques
• Experience working with large datasets or fine-grained classification problems
• Strong programming skills in Python and/or C++

Nice to Have:

• Experience with image retrieval / similarity search / metric learning
• Familiarity with automotive domain or parts catalogs
• Experience with multimodal systems (image + metadata measurements)
• Experience working in early-stage or R&D-heavy environments

Critical Soft Skills

Problem-Solving & Adaptability

• Ability to work on complex, ambiguous problems without predefined solutions
• Willingness to experiment, fail, and iterate quickly
• Capability to identify when an approach is not working and pivot in time

Pragmatism & Decision Making

• Focus on practical, working solutions, not just theoretical models
• Ability to balance accuracy, cost, and implementation complexity
• Comfortable making decisions under uncertainty with limited data

Ownership & Accountability

• Takes full ownership of outcomes, not just tasks
• Drives progress independently in an R&D environment
• Proactively identifies risks and improvement areas

Communication

• Ability to explain technical concepts and trade-offs in clear, business-oriented language
• Works effectively with cross-functional teams (engineering, product, operations)
• English level: Upper-Intermediate

Responsibilities:

AI & Model Development

• Design and implement computer vision models for automotive parts recognition
• Develop embedding-based and/or classification models for large-scale part catalogs
• Build and maintain training, validation, and evaluation pipelines
• Optimize models for real-world performance (mobile images, noise, occlusions, lighting variability)

Data & Experimentation

• Define dataset requirements and contribute to data collection strategy
• Work with both controlled (studio) and real-world (field) image datasets
• Apply augmentation, domain adaptation, and data balancing techniques
• Continuously evaluate model performance and iterate based on results

Problem Solving & Architecture

• Identify limitations of current approaches and propose alternative solutions
• Design hybrid approaches (e.g., visual recognition + metadata or measurements)
• Handle edge cases such as visually similar parts and ambiguous inputs
• Make trade-offs between accuracy, performance, and implementation complexity

Integration & Collaboration

• Collaborate with mobile and backend engineers to integrate models into the application
• Support deployment and optimization of models for production environments
• Communicate technical decisions, risks, and limitations to stakeholders

Continuous Improvement

• Improve model performance over time using newly collected field data
• Contribute to retraining and dataset expansion strategy
• Monitor real-world performance and identify improvement opportunities