Senior Machine Learning Engineer

The role

You will design, build, and operate model-serving and inference infrastructure for transformer-based vision and multimodal models. You will own the serving architecture and optimize models and pipelines for production.

What you will do

• Design, build and operate model-serving and inference infrastructure for transformer-based vision and multimodal models
• Own the serving architecture behind our ML features
• Optimise models and pipelines for production through batching, KV caching, GPU utilisation and orchestration
• Integrate models into Encord's product and annotation workflows
• Build reusable abstractions that work across models, data types, customers and changing task templates
• Set up monitoring and observability
• Take prototypes from research-focused teammates and turn them into shipped features
• Contribute to platform design decisions, take part in design reviews and help us hire

Required qualifications

• 5+ years of software engineering experience, including meaningful time building and operating ML systems or model-backed products in production at scale
• Strong software engineering and system design fundamentals
• Hands-on experience deploying modern deep learning models in production, with a strong understanding of model serving, inference optimisation, GPU utilisation and performance trade-offs
• Experience designing and operating scalable, distributed production systems, ideally involving GPU workloads, high-throughput data processing or low-latency inference
• The ability to reason through an ML system end to end, from data and model choices to offline and online evaluation, deployment, monitoring and retraining
• A track record of personally owning complex systems from design through to production
• Work with Python, PyTorch, CUDA, Kubernetes and GCP/AWS

Nice to have

• computer vision or multimodal deployment experience
• model-serving frameworks
• experience with video, 3D or other complex data types

Benefits

• Competitive salary
• meaningful equity in a high-growth startup
• 25 days annual leave + UK public holidays
• Annual learning & development budget
• Travel for customer visits, events, and conferences across the UK and Europe
• Daily company lunches
• Monthly socials & bi-annual team offsites

About the company

Encord is the universal data layer for AI that helps 300+ AI teams train and run models on the right data. They are an ambitious team of 100+ working at the frontier of AI.