All open roles

Computer vision

Computer Vision Engineer

Location
Milford, Derbyshire
Type
Full-time
Working pattern
Hybrid: 2 days on-site, 3 remote

We're building up our computer vision team and looking for a Computer Vision Engineer. You'll own the models and infrastructure we already have running on real footage of animals: keeping them reliable, measuring how they perform in the field, and steadily improving them. There is room to build new things too, but the first job is making sure what we've shipped keeps working and gets better.

What you'll do

  • Maintain and improve the computer vision models we have in production (detection, tracking, behaviour classification) on images and videos from real environments
  • Own the training and inference infrastructure: data pipelines, retraining, versioning, and deployment to cloud and edge devices
  • Monitor model performance in the field, investigate what breaks (lighting, occlusion, camera placement, seasonal change) and engineer around it
  • Build the test sets and evaluation that tell us when a change is safe to ship
  • Work with our annotation specialists to shape labelling guidelines, spot data quality problems early, and close the loop between model failures and the data that fixes them
  • Support new model work as it moves from experiment to production
  • Input on the technical direction of the vision pipeline and the trade-offs between accuracy, cost and deployment constraints

What we're looking for

  • 3+ years in computer vision: with real production experience, not just research or prototyping
  • A relevant degree (Computer Science, Maths, Engineering, or similar)
  • Shipped computer vision work: you've taken models from training into something people actually used, and you can talk about what went wrong
  • Strong Python: clean, tested code; comfortable with the full training and evaluation loop
  • Video and image fundamentals: object detection, tracking and/or segmentation; OpenCV; a working grasp of what cameras, compression and lighting do to your data
  • Evaluation judgement: you know the difference between a good and a bad model, and you build test sets that expose the gap
  • Deployment experience: optimising and deploying models to constrained hardware, edge devices or cloud pipelines (ONNX, TensorRT, quantisation, or similar)
  • Collaborative, honest and direct: you say what you think in review, take feedback, and follow through on what you commit to
  • Care about the mission: an interest in using tech for real-world animal welfare impact

Nice to have

  • Multi-object tracking or behaviour recognition in video
  • Edge or embedded deployment (Jetson, Coral, or similar)
  • Experience designing data strategy: what to label, how much, and when you have enough
  • Experience with agricultural imagery or a related field
  • An interest in agriculture, animals or veterinary science

What we offer

  • Hybrid working: 2 days/week on-site in Milford, 3 days remote
  • Real ownership from day one, and a clear path to a senior role
  • A team where your decisions ship and you see them running on sites
  • Meaningful work: your models directly improve animal welfare

Why this role matters

A model that worked in March and quietly degrades by January is a welfare failure nobody notices until it's too late. Your work is what stops that happening. Keep our systems reliable and improving and animals get help earlier and live better lives.

If you want your computer vision work to run somewhere real and count for something, we'd love to hear from you.

About Vet Vision AI

Vet Vision AI is a University of Nottingham spin-out, founded by veterinary surgeons on a simple belief: animals are always telling us how they are, just not in words, but through their behaviour. We build vet-led, welfare-focused computer vision that watches continuously and turns behaviour into clear, actionable welfare evidence for dairy and equine teams, helping them deliver great care.

Our systems are already deployed on commercial farms and equine yards and trusted within the supply chains of major retailers. We are a small, fast-moving, science-led team on an enormous mission: to revolutionise the welfare of millions of animals worldwide, and we are just getting started.

How we work

We're an agile team: clear priorities, honest reviews, and retros that actually change how we work. We're collaborative by default and encourage everyone, across every department, to work closely together and learn from one another.

Our location

We're based in Milford, Derbyshire, a village on the edge of the Peak District with strong transport links to Derby, Nottingham and the wider Midlands area. The office is a short walk from the River Derwent and a few minutes from the hills, so lunchtime walks are part of the rhythm here.

5 Step Hiring Process.

You'll meet the people you'd actually work with, talk through the role properly, and do a practical exercise drawn from the real job. We'll tell you what each step is upfront, and you'll hear from us quickly, either way!

01

Apply or introduce yourself

Send your CV, LinkedIn, portfolio, or a short note on why this work matters to you.

02

Introductory culture call

A short conversation about you, the role and our culture at Vet Vision AI.

03

Hiring manager interview

We talk through the role in detail: the good, the challenging, and the rewarding.

04

Practical task

A realistic exercise close to the actual job.

05

Final co-founder interview

This is a great chance to pick Charlie and Robert's brains and for them to challenge yours!