AI and Computer Vision Intern; R&D

Our mission is to build transparent and sustainable business relationships with our suppliers by providing exceptional service, secure transactions, and a clear understanding of the value of their materials.

Since its founding in 1996, PMR has established itself as a leader in the recycling and recovery of precious metals contained in automotive catalytic converters. Through our commitment to innovation, integrity, and operational excellence, we have built a strong reputation across North America.

Today, PMR operates 3 processing plants and serves a diverse client base, including automotive recyclers, metal recyclers, catalytic converter buyers, industrial catalytic converter recyclers, and metal brokers.

Our expertise is based on a unique processing method that maximizes the recovery of precious metals such as platinum, palladium, and rhodium. Since the concentration of these metals varies significantly from one catalytic converter model to another, precise part identification is a key strategic factor in the profitability of our operations.

In a context of growth and continuous innovation, we are actively investing in the development of new technologies, automation, and artificial intelligence to improve our processes and maintain our position as an industry leader.

Catalytic converter identification is currently performed by experts. The goal of this internship is to automate all or part of this process using artificial intelligence, computer vision, and advanced statistical models.

For over a year, we have been developing and testing various approaches (computer vision, deep learning, heuristics, hybrid models). The intern’s role will be to explore new technical directions, experiment with unconventional approaches, and propose solutions that can be combined with our existing systems to improve performance.

We primarily use open-source technologies.

By the end of the internship, you will have:

• Worked on a real, complex industrial problem
• Developed concrete and verifiable skills in computer vision and applied AI
• Contributed to solutions at the cutting edge of the state of the art, well beyond academic tutorials

We are looking for a curious, rigorous, and creative individual, motivated by complex challenges in computer vision and artificial intelligence. Simple solutions have already been explored—this internship is for someone who enjoys digging deeper, testing, failing, and iterating.

We already have a list of potential directions for you to explore as a starting point, but we expect you to identify additional avenues as well.

You must:

• Be able (with the help of AI tools if needed) to program in Python or Ruby
• Have a good understanding of computer vision models (e.g., YOLO, Transformers, CNNs, etc.), sufficient to use, fine-tune, and potentially modify them
• Be comfortable working in a Mac and/or Linux environment
• Have an analytical mindset and a scientific approach to experimental results
• Interest in model optimization, multimodal models, or hybrid approaches (vision + statistics) is an asset

• Remote work possibilities. Minimum of 3 days per week required at our Boisbriand office
• Starting salary of $23/hour, paid weekly
• Close collaboration with a senior developer, providing strong technical mentorship while allowing significant freedom in methodological choices
• Free parking and coffee