Performance Engineer - Containers, Serverless & ML GPUs

Verda; Greater London; Hybrid; GBP 77,000 - 128,000; Full time

Benefits offered by this job

Healthcare

Lunch

Wellbeing

Equity

Job summary

Verda is seeking a Performance Engineer to own the end-to-end surface for its container and serverless GPU platforms. You will profile, optimize, and drive improvements across image distribution, startup, and data loading paths to reduce cold-start latency and boost throughput.

You will work across storage, compute, networking, and platform teams to remove bottlenecks and publish insights that inform internal and customer-facing decisions.

Qualifications

• Must have demonstrable experience making ML/AI workloads faster in production environments.
• Strong understanding of Linux storage and container runtime internals.
• Hands-on experience with distributed or network filesystems and S3-compatible storage, incl. performance pitfalls.
• Familiarity with model-serving runtimes (e.g., vLLM, SGLang) and formats (safetensors, GGUF, checkpoints).
• Ability to reason about NIC, switches, filesystem, cache, container runtime, and GPU as one system.

Responsibilities

• Profile and optimize end-to-end path for containerized ML/AI workloads.
• Design and tune storage between object stores and GPU nodes.
• Drive measurable wins in latency and throughput across services and workloads.
• Benchmark real workloads and translate findings into platform changes.
• Collaborate across compute, networking, and platform teams to remove bottlenecks.
• Publish internal write-ups to share trade-offs and findings.

Skills

ML/AI performance

Linux storage internals

Distributed storage

S3-compatible storage

Model-serving runtimes

End-to-end systems view

Job description

Verda is seeking a Performance Engineer to own the end-to-end surface for its container and serverless GPU platforms. You will profile, optimize, and drive improvements across image distribution, startup, and data loading paths to reduce cold-start latency and boost throughput.

You will work across storage, compute, networking, and platform teams to remove bottlenecks and publish insights that inform internal and customer-facing decisions.