1 day ago
London, United KingdomStaff+
Responsibilities
- Build and prototype ML platform features, workflows, SDKs, APIs, and developer tooling.
- Develop reference implementations for training, fine-tuning, inference, model serving, and orchestration.
- Integrate ML frameworks and tools with GPU, Kubernetes, storage, networking, and platform services.
- Partner with Product, engineering, infrastructure, and customers to identify friction and improve developer experience and time-to-value.
- Evaluate emerging ML infrastructure technologies and support product discovery through prototypes, benchmarks, demos, and proof-of-concepts.
- Define technical requirements, documentation, examples, and best practices for ML services.
Requirements
- Strong engineering experience, ideally in ML infrastructure, developer tooling, cloud platforms, or AI products.
- Strong understanding of ML workloads including training, fine-tuning, inference, and model serving.
- Understanding of GPUs, Kubernetes, containers, storage, networking, APIs, SDKs, CLIs, notebooks, and distributed systems.
- Familiarity with Ray, Slurm, MLflow, Hugging Face, vLLM, Triton, LangChain, or similar technologies.
- Ability to move from ambiguous customer problems to working prototypes or production implementations.
- Strong communication skills and comfort working directly with product teams and customers.
Benefits
- 25 days of annual leave.
- Private medical insurance through Bupa.
- Cycle to Work Scheme.
- Gympass subscription to gyms and wellbeing apps.
- Participation in the company shares program.
- Enhanced parental pay and leave.
- Dedicated learning time and a collaborative culture with regular feedback.
- The posting does not state a work arrangement or contract duration.
Tech Stack
KubernetesMLflow
Categories
About Radiant
RadiAnt develops RadiAnt DICOM Viewer, a lightweight Windows application for viewing and analyzing medical images from CT, MRI, PET, ultrasound, and X‑ray. It sells commercial licenses (with a free trial) to radiology practices, hospitals, and researchers, and supports PACS query/retrieve, multi-modality workflows, and 2D/3D reconstructions. The software can open studies from local media and networked DICOM servers for clinical review and teaching use cases.