Applied AI Scientist: On-Device Vision & LLM Systems

Popsa Group; Greater London; Hybrid; GBP 100,000 - 150,000; Full time

Job summary

Popsa in London is hiring an experienced ML specialist to own end-to-end projects from brief to production. You will train, distill and quantise models that run on device and in our FastAPI services, collaborating with iOS and Android engineers.

You'll define success criteria, evaluate large image models, and communicate results to product and leadership. This is an office-first role with hands-on experimentation and rapid delivery.

Qualifications

• Hands-on experience training deep learning models on large image datasets, incl. transfer learning and fine-tuning.
• Proven end-to-end ML project ownership from brief to deployed model.
• Rigorous evaluation with chosen metrics, datasets and thresholds; document results.
• Strong Python skills and ability to read/extend existing code.
• Docker and CI/CD in normal workflow; Terraform familiarity.
• Excellent communication of design notes and PR descriptions.
• Solid grounding in statistics and ML fundamentals.
• Experience with AWS SageMaker, S3, ECS or similar.
• Shipping models on device and integrating with mobile engineers is ideal.

Responsibilities

• Own projects end to end: take an undefined brief, break it into pieces, and drive them through training, evaluation, deployment and monitoring.
• Define success metrics and evaluation sets with stakeholders up front.
• Collaborate with mobile engineers to ship models on device and to framework.
• Design, train and deploy deep learning models that reach customers on device and server.
• Keep work visible with updates and tickets that move, sharing intermediate results.
• Stay current on ML research and tooling and bring useful ideas to the team.
• Improve infrastructure for large-scale training, evaluation and experiment tracking.

Skills

Deep learning

Transfer learning

Fine-tuning

Distillation

Quantisation

End-to-end ML projects

Evaluation metrics

Python

Writing & communication

Statistics & ML fundamentals

AWS SageMaker

Tools

Docker

CI/CD

Terraform

SageMaker

Job description

Popsa in London is hiring an experienced ML specialist to own end-to-end projects from brief to production. You will train, distill and quantise models that run on device and in our FastAPI services, collaborating with iOS and Android engineers.

You'll define success criteria, evaluate large image models, and communicate results to product and leadership. This is an office-first role with hands-on experimentation and rapid delivery.