AI Infrastructure Engineer

mercor.com; San Francisco (CA); Vor Ort; USD 180.000 - 240.000; Vollzeit

Benefits dieser Stelle

Performance bonus

Equity grant

Relocation bonus

Housing stipend

Meals stipend

Equinox membership

Laundry reimbursement

Wellness reimbursement

Health insurance

Dental insurance

Vision insurance

Zusammenfassung

Mercor is seeking an experienced software engineer to own the accelerator platform end-to-end. You will manage GPU clusters across cloud and on‑prem, write production code daily, and design systems that keep accelerators busy and highly utilized.

You will work on scheduling, LLM serving optimization, benchmarking and cost attribution, with on‑call responsibilities. This is a hands-on, production‑oriented role in San Francisco offices.

Qualifikationen

• 5+ years building and operating production distributed systems or ML infrastructure.
• Proficient Python and Go or C++, with shell scripting.
• Strong Kubernetes and HPC experience with Slurm.
• Experience with distributed training or serving stacks (PyTorch Distributed, Ray, vLLM, Megatron-LM).
• Experience with at least one major cloud provider.
• Familiar with observability tools and on-call duties.

Aufgaben

• Design, deploy and operate GPU clusters across AWS, GCP, and on‑prem data centers.
• Build scheduling to keep accelerators busy: gang scheduling, bin-packing, multi‑tenant quotas.
• Run distributed training infra with PyTorch Distributed, Ray or Megatron.
• Deploy and tune high‑throughput LLM serving with vLLM and quantization.
• Build benchmarking and traffic‑replay tools to measure latency & baselines.
• Health checks and automated remediation across hardware, network and software.
• Instrument fleet utilization with DCGM, Prometheus and Grafana; write alerts.
• Ingest utilization and billing data to support capacity planning.
• Drive autoscaling and idle-capacity recovery with measurable cost results.
• Write runbooks and explain infra tradeoffs to researchers and leadership.

Jobbeschreibung

ABOUT MERCOR

Mercor's mission is to organize human intelligence to power the AI economy. We're a leading AI data company, building the layer between human expertise and frontier models. Millions of domain experts on the platform are paid over $4 million per day to train frontier AI models. Mercor's APEX benchmark family measures AI's real-world impact on professional work. Mercor Enterprise brings this same infrastructure to Fortune 500 companies: helping companies capture how their best people actually work, translating that expertise directly back into agents.

Mercor is creating a new category of work where expertise powers AI advancement. Achieving this requires an ambitious, fast-paced and deeply committed team. You’ll work alongside researchers, operators, and AI companies at the forefront of shaping the systems that are redefining society. Mercor is a profitable Series C company valued at $10 billion. We work in-person five days a week in our San Francisco, NYC, or London offices.

Summary of role

All of Mercor’s AI work, from training and serving models to building agents, runs on GPUs. Your job is to make that compute fast, fault-tolerant and fully utilized.

You’ll own the accelerator platform end to end. That means the clusters and schedulers across AWS, GCP, neoclouds and on-prem, the training and LLM serving on top, and the observability and cost data that show where every GPU-hour goes.

This is a software engineering role, not an ops role. You’ll write production code every day, own systems from design through on-call, and build the platform that sets how fast Mercor can move.

Representative projects

• Stand up a multi-cloud GPU fleet spanning hyperscalers, neoclouds and on-prem, orchestrated through custom Kubernetes and Slurm operators so researchers never have to think about where their job lands.
• Build a scheduling layer with gang scheduling, bin-packing and fair-share quotas that measurably raises accelerator utilization across teams.
• Ship a high-throughput vLLM serving stack with prefill/decode disaggregation, KV cache offload and FP8/FP4 quantization. Back it with a benchmarking and traffic-replay harness that tracks TTFT, TPOT and latency percentiles against per-config baselines.
• Build a self-healing health-check and remediation system that rides out fault storms across hardware, network and software without paging a human.
• Build an end-to-end cost-attribution pipeline. It ingests utilization, occupancy and billing data into the warehouse, attributes every GPU-hour to a team and workload, and drives capacity planning.
• Build internal tools and dashboards that researchers and non-engineers use every day.

Responsibilities

• Design, deploy and operate GPU clusters across AWS, GCP, and other hyperscalers/neoclouds/on-premise data centers, including custom Kubernetes and Slurm operators and the metadata services behind them.
• Build scheduling that keeps accelerators busy: gang scheduling for distributed jobs, bin-packing to reduce fragmentation, and quotas and fair share for multi-tenant teams.
• Run distributed training infrastructure on PyTorch Distributed, Ray, or Megatron, with fast checkpointing and model loading.
• Deploy and tune high-throughput LLM serving with vLLM or similar: tensor and model/expert parallelism, prefill/decode disaggregation, KV cache offload, FP8/FP4 quantization.
• Build benchmarking and traffic-replay tools that measure TTFT, TPOT and latency percentiles, and set per-config performance baselines.
• Build health checks across hardware, network and software layers, and automated remediation that handles fault storms without human intervention.
• Instrument fleet utilization with DCGM, Prometheus and Grafana. Write recording rules and alerts that teams rely on.
• Ingest utilization, occupancy and billing data into a warehouse, attribute spend to teams and workloads, and support capacity planning.
• Drive autoscaling, rightsizing and idle-capacity recovery programs with measurable cost results.
• Write clear runbooks and documentation, and explain infrastructure tradeoffs to researchers, finance and leadership.
• Build internal tools or dashboards that researchers and non-engineers use daily.

You may be a good fit if you

• Have 5+ years building and operating production distributed systems or ML infrastructure
• Write production-quality Python plus Go or C++, and are fluent in shell scripting for system administration.
• Have deep hands-on Kubernetes and container experience, plus working knowledge of Slurm or another HPC resource manager.
• Have run GPU workloads at scale, including scheduling, orchestration and troubleshooting.
• Have hands-on experience with at least one distributed training or serving stack: PyTorch Distributed, Ray, vLLM or Megatron-LM.
• Have deep experience with at least one major cloud provider.
• Know observability tools like DataDog, Prometheus, PromQL and Grafana, and are comfortable on-call for critical systems.
• Own projects end to end and gather your own requirements with little direction.

Strong candidates may also have experience with (or a strong desire to learn deeply)

• LLM inference optimization: KV cache systems, disaggregated serving, quantization, speculative decoding
• High-performance networking: RDMA, InfiniBand, NCCL, NVLink
• GPU profiling with Nsight or similar, and CUDA or Triton kernel work
• Accelerators beyond NVIDIA: AMD MI-series, TPU, Trainium
• Capacity planning, billing normalization or cost attribution across providers
• Automated fault diagnosis, log mining or AIOps at fleet scale
• Configuration management (Ansible, Puppet) and secure or air-gapped deployments

Benefits

• Bi-annual performance bonus structure
• Generous equity grant vested over 4 years
• Up to $15k Relocation bonus
• $10K housing bonus (if you live within 0.5 miles of our office)
• $1.5K monthly stipend for meals
• Free Equinox membership
• $200 monthly laundry reimbursement
• $200 monthly personal wellness reimbursement
• Health, Dental, Vision insurance