About Captur
Captur helps software understand real world scenes in real-time with an SDK for flexible, on-demand visual recognition. We’re a small, rapidly scaling team backed by top-tier investors; we recently closed a $6M seed round to accelerate product and go-to-market growth. We are global leaders in edge ML and have validated +150M images on-device for enterprise customers such as Lime. Next, expanding as a horizontal platform across use cases that require real-time speed, high volume and coverage across a wide range of mobile devices. 

About the role

You’ll work directly with our ML team on a focused project that aligns both with your interests and our product roadmap. Throughout the internship, you’ll receive regular mentorship and technical feedback while working on real machine learning challenges involving models that run directly on users’ devices.

Potential project areas include:

• Optimising computer vision models for efficient inference on mobile devices
• Fine-tuning and deploying multimodal vision-language models for real-world scene understanding on the edge
• Developing data-centric ML systems that leverage automated labelling and synthetic data generation to improve model quality
• Improving model robustness across challenging real-world environments, edge cases, and distribution shifts

What We’re Looking For

Required:

• We are looking for candidates who are currently pursuing, or have recently completed, a Master’s or PhD in Computer Science, Machine Learning, Mathematics, Engineering, Physics, or a related quantitative discipline
• Strong Python programming skills
• Demonstrated machine learning experience through research, internships, coursework, or substantial personal projects
• Experience independently delivering a technical project from problem definition through experimentation and evaluation
• Clear written and verbal communication

Helpful but not required:

• PyTorch, TensorFlow, or JAX experience
• Computer vision or multimodal ML experience
• Experience with model optimisation or edge deployment
• Experience building ML infrastructure, evaluation systems, or data pipelines

We’re particularly interested in candidates who enjoy understanding how ML systems behave in production, not just training models.

Role Compensation and Working Details

• 8 - 12 week internship (flexible start date)
• Base salary of £35,000 per annum, pro-rata'd for the duration of the internship.
• Taxable housing stipend of £125 per week for interns whose permanent address is outside London and who need to pay for accommodation during the internship.
• Based 3 days a week from our Liverpool Street Office, with work from home on the remaining days.
• 25 days' holiday plus public holidays, pro-rata'd for the duration of the internship
• Dedicated company Macbook Pro for use during the internship