Machine Learning Operations

MLOps Engineer A venture-backed deep tech startup is building an AI accelerator that computes with light. Role Overview As the first dedicated MLOps Engineer, you will own the infrastructure that carries models from research to silicon-validated production: the pipelines, tooling and platforms that let the AI and hardware teams move quickly while the chip and its software stack are still being built in parallel. It is a high-ownership role sitting between ML research, compiler stacks and novel hardware. What You'll Own Pipelines and Deployment * Designing and operating end-to-end ML pipelines: data ingest, training, evaluation, quantisation and deployment onto custom accelerator hardware * Building experiment tracking, model registry and versioning infrastructure (MLflow, W\&B or equivalent) tuned to hardware-in-the-loop workflows * Instrumenting and monitoring production inference deployments, with alerting and rollback strategies suited to hardware-accelerated serving Testing and Benchmarking * Owning CI/CD for ML: automated testing of model correctness, numerical accuracy and on-chip performance after every change to models, compilers or firmware * Building tooling to benchmark inference on custom silicon across latency, throughput, power and utilisation Infrastructure and Platform * Managing compute scheduling across on-premises accelerator clusters and cloud GPU/CPU for training and simulation workloads * Driving infrastructure-as-code: containerisation, orchestration (Kubernetes or Slurm) and reproducible environment management * Building the internal developer platform: self-service tooling, documentation and runbooks that lift engineering productivity * Working with ML researchers, compiler engineers and hardware architects to find and remove bottlenecks across the model-to-chip workflow What We're Looking For Essential * 5+ years in software or infrastructure engineering, with at least 2 in an ML or AI-adjacent role * Strong Python and familiarity with PyTorch or JAX, comfortable reading and modifying model code * Hands-on experience building and operating production ML pipelines: data, training orchestration, evaluation and serving * Experience with experiment tracking and model lifecycle tools (MLflow, W\&B, DVC or similar) * Solid grasp of containerisation (Docker) and orchestration (Kubernetes or Slurm) for distributed compute * Infrastructure-as-code (Terraform, Ansible or equivalent) and CI/CD (GitHub Actions, Jenkins or similar) * Experience with hardware-accelerated compute (CUDA/GPU workflows, profiling, performance tuning), even if not on custom silicon * Strong observability and debugging: distributed tracing, logging, metrics dashboards * Comfortable working where the hardware and software are being built at the same time Useful * Experience with custom or novel accelerators (FPGAs, ASICs, NPUs or research chips) * Familiarity with ML compiler stacks: MLIR, LLVM, TVM, XLA or vendor compilers * Model optimisation: quantisation (INT8/INT4/FP8), pruning, distillation or mixed-precision training * On-chip performance profiling and roofline analysis * Chip bring-up: running early software stacks on pre-silicon simulation or first silicon * Open-source contributions to ML infrastructure or compiler tooling * Background in deeptech, semiconductor or hardware startups What's On Offer * A genuine technical challenge: owning the ML infrastructure behind an accelerator that computes with light, as the first dedicated MLOps hire * Share option scheme, so you share in what you build * Competitive salary * Private health insurance (AXA) and pension (Aviva) * £500 annual learning and development budget * 25 days holiday plus bank holidays, subsidised on-site lunches, cycle-to-work and regular company socials Get in touch for a confidential conversation. Imogen@waverecruitment.co.uk