Want to know if this job is worth applying to?
Greater London, England, United Kingdom
Hybrid
Location: London (Hybrid), Safi HQ in Spitalfields
Salary: £40,000 – £80,000 + equity
Travel: Regular site visits across Europe, North America and Latin America
Safi's mission is to make circular economy firms more profitable through the deployment of AI technology. We do that by developing foundational AI models, software, and data connectors.
As an AI Product Manager for Vision Models, you are the critical layer that connects on-the-ground AI performance with our AI/ML engineering.
You will:
Capture and triage data + feedback from customer deployments
Use your judgement and understanding of our customer's industrial processes + priorities, recycled material, and our vision AI capabilities to prepare and prioritise work for our AI/ML engineers
Lead the domain discovery required to bring new materials and markets online
Own the data pipeline and labelling teams that we use to teach our models
You will spend time at recycling sites watching how operators actually assess quality - and translate what you see into precise AI model development, labelling and training sets. You must be comfortable planning with ML engineers, reading model-performance data, and reasoning about trade-offs with people.
We think much of this role can be learnt on the job - high energy, commitment, quick thinking and willingness to get your hands dirty (quite literally, we work with waste materials) will go far. This is a junior-mid level role for someone who is excited about deploying AI that impacts real-life industrial plants.
You do not need prior experience as a product manager in a tech company.
Our customers are industrial recyclers of plastic and metals — manufacturers, processors, smelters. These firms are held back by limited, legacy technology and fragmented data. Within our first year of operations, we have signed customers such as one of the world's largest recycling plants, a top 5 global aluminium smelter and a group that processes the entire plastic waste stream of a major nation.
We have product-market fit and strong customer traction in an under-served market. We want to expand on that traction to help manage the entire end-to-end lifecycle of plants in multiple sectors.
We're backed by leading climate-focused VCs, including LowerCarbon Capital, Nosara Capital, and Transition Ventures. If our mission resonates with you, we encourage you to apply, even if your experience doesn't match every requirement.
Own the ML backlog and prioritisation. Triage incoming data and feedback from customers and colleagues. Assess with sales, deployment and customers to understand importance. Execute the work - training preparation, data collection, labelling - that our ML engineers need to solve the customer problems
Review model feedback and spot trends. Regularly review incoming AI-model feedback, identify recurring issues or trends, and feed the significant ones into the ML triage process.
Maintain and curate the labelling guidelines (taxonomy). Keep the guidelines accurate and up to date as understanding improves and new edge cases emerge. Decide when to split out a new class and when to fold classes together — decisions that directly determine final model quality.
Lead new-material and new-partner model development. Understand deeply with our customers how different objects, contamination and material affects their process. Understand what matters when assessing its quality: what can be judged visually, how accurate detection needs to be, and where the hard cases are.
Gather detailed specifications, convey them to the ML team, sanity-check feasibility, create the labelling guidelines, and get annotators onto the work. Run weekly reviews with partners.
Act as the interface between the business and the ML team. Run the weekly cadence with ML engineers, deployment, and sales so that priorities, guidelines, and progress stay aligned.
ML literacy: you have a basic understanding of how computer vision AI works
Data analysis: you have experience analysing data and taking accurate conclusions
Great people skills and confidence: You will need to walk onto an industrial site and be able to connect with forklift operators and plant managers, some of whom will not be able to speak the same language as you
Information absorption: you will need to be able to quickly absorb information about materials, business needs, sales requirements and make a plan
Able to write and maintain detailed documentation for your team. This is the core of taxonomy management.
Genuine curiosity about the physical material and how it's handled
Resilience: willing to travel extensively to recycling and processing sites in un-glamorous parts of the world
Experience in a ML/data product role — ideally working directly with a machine-learning, data-science, or computer-vision team
Experience in recycling, waste, materials, manufacturing, or another physical/industrial domain.
Working knowledge of Spanish
We're a small team and we all automate our own work - most people here use LLM and agentic tools to streamline triage, reporting, and recurring processes. We use Linear, Notion, Claude, GitHub and GCP. If you have a builder's instinct for making your own job easier, you'll fit in well.
We're in our office near Spitalfields / Brick Lane a few days a week for the energy of building together in person. You will also be travelling 1-2 times per month.
Salary between £40,000 and £80,000
Share option plan - every employee owns what we're building
26 days annual leave (+ all UK bank holidays) - the bank holidays are flexible, so you can take them whenever it suits you
Personal wellness & development budget of £75 per month
Home office kit-out budget of £500
Private health insurance (opt in)
Business and leisure travel insurance
Salary sacrifice pension scheme
Cycle to work scheme
Safi is an equal opportunity employer. We welcome applicants from all backgrounds and do not discriminate on the basis of race, colour, religion, sex, sexual orientation, gender identity, national origin, age, disability, or any other status protected by applicable law. We're committed to building a diverse team and making our hiring process fair and accessible. Please let us know if you need any reasonable adjustments during the recruitment process.