Machine Learning Engineer

A little bit about us  

At Recycleye we are revolutionising recycling. Our ground-breaking AI-powered robots and optical sorters are making waste sorting faster, cleaner and more profitable. With our advanced tech, the previously unsortable can now be sorted efficiently and effectively. Founded in 2019, we have grown rapidly ever since, with installations at leading waste management companies across Europe and the US. 

We are on a mission to help our clients produce higher quality outputs - turning the world’s waste into valuable resource. And our work has never been needed more. Worldwide only 9% of plastic waste is being recycled - every day, we are proud to work on making a meaningful difference towards a sustainable future. 

Be part of it - come and join us!  

Why choose to work at Recycleye? Our team works together to grow and succeed, driven by our shared values of ambition, curiosity, customer centricity, integrity and respectful ideation. We offer great benefits, built-in upskilling and unlimited holiday. 

We know that our mission will be achieved by having the most passionate, driven, talented, and all-around best people on our team. So come and help us shape the future! 

What you'llwork on  

Our team develops new features across Recycleye’s waste sorting and plant monitoring products, improving the end-to-end performance of our systems. That might mean evaluating and deploying new model architectures, adapting ideas from ML papers to our use cases, implementing novel object tracking algorithms, diving deep into camera optics and calibration, or making our DeepStream pipelines more efficient. Everything we build runs in production on real sorting lines, so we look for the solution that works best for the problem, which might mean stepping back and choosing a better camera rather than chasing a cool new model. 

You will be a core member of a small, high-impact ML R&D team. Building new features and developing better ways to measure how well our products perform, whether that’s sorting a waste stream or understanding what’s in it, and how real site conditions affect that performance. You will work closely with our data team, who train models day to day for client projects and rely on the tools we build, and with colleagues developing a new plant monitoring product. You will also help support our deployed systems and spend time on site, getting hands-on with the waste they sort. 

As the team develops, you will help shape how we work, from code review to testing, and take on more responsibility for planning and decision-making. Along the way you will build a deep understanding of our products and the waste sorting industry, keep up with advances in computer vision, and share what we learn across the company. 

We work mostly in Python, with some Go, deploying models and algorithms to NVIDIA Jetson devices running DeepStream pipelines, supported by Docker, GitHub Actions and Azure. 

What success looks like  

• Your work regularly reaches production and makes a measurable difference to how our products perform. 

• You bring new ideas to the table, test them, and help shape what we build next. 

• You grow into owning technical areas of our systems, and become a trusted reviewer and planning partner for the team. 

• Teams across the company, from data to install and support, know to come to you, and you help them get the best from our systems. 

• Our ways of working and measuring performance keep improving, so that our systems get better with every deployment. 

What you'llbring  

We’re looking for someone who is proactive, driven and practical, comfortable with change, and keen to learn. We don’t expect you to be an expert in everything on day one. 

• Hands-on computer vision experience, with a background in ML, physics, maths, engineering or computer science. 

• Around 3-5 years’ experience writing production-quality code, primarily in Python, and promoting good engineering practice. 

• A track record of shipping ML into production in a commercial setting, ideally including computer vision on edge devices. 

• Solid software engineering fundamentals: tested code, version control and code review. 

• Clear communication with both technical and non-technical colleagues, and a collaborative approach. 

• Comfort working independently and as part of a small team. 

• Creative, practical problem solving. 

• A love of seeing your work in the real world. You’ll spend time on site and at our demo centre, testing in challenging environments and getting hands-on with the waste our systems sort. It’s the best way to understand the problem you’re solving. 

Nice to have  

We don’t expect you to have all of these, and we’ll help you build the rest: 

• Models: computer vision detector-classifier models (YOLO, RF-DETR), segmentation models 

• Algorithms: multi-object tracking, camera calibration, homography transforms 

• Edge deployment: NVIDIA Jetson, DeepStream / GStreamer 

• Infrastructure: Docker, CI/CD (e.g. GitHub Actions), cloud platforms (Azure, AWS, GCP, etc.) 

• MLOps: model evaluation and monitoring in production 

• Programming languages: Go 

• Tooling: building tools used by other teams 

• Hardware: physical systems or robotics, cameras and lenses 

• Commercial awareness: understanding how technical choices affect customers and the business 

What we offer   

• Day 1 health insurance through Vitality  

• Unlimited holiday 

• £1,000 professional development budget per year  

• Monthly team socials  

• Direct influence over technical direction 

Practicalities  

Location: Stratford E15 2HU, Hybrid working  

Travel: Occasional to Europe or US 

Even if you don’t tick every box listed above but are motivated by our work and believe you can make a difference on our team, we’d really love to hear from you - so please do apply and drop us a note about what you bring!