The Roster.

The binding constraint on AI scaling has shifted to powered land itself — firms now hire dedicated acquisition specialists to secure gigawatt-capable brownfield sites.

$200,000 – $238,000

Datacenter site selection is now a $200K+ specialised real-estate-plus-power-engineering job — sourcing 200MW+ behind-the-meter power at a single site is its own function, not generic CRE.

$200,000 – $238,000

Member of Technical Staff: Research

Zep · Research

San Francisco, CA / Remote (US)

Enterprise buyers are paying for agent memory as its own infrastructure layer, not a RAG bolt-on — personalised agents at scale is now a standalone research problem at lead-level comp.

$180,000 – $250,000 + 1–1.5% equity

Network Engineer — ML Infrastructure (High-Speed Interconnects)

xAI · Infrastructure

Palo Alto, CA

Hiring for copper and optical interconnect fabric specifically for training and inference clusters confirms interconnect, not just GPU count, is a named bottleneck at xAI's current scale.

$180,000 – $440,000

Member of Technical Staff — RL Inference

xAI · Research

Palo Alto, CA

This role exists to optimise the inference stack for RL training workloads — RL rollout has become its own bottleneck, distinct from serving inference.

$180,000 – $440,000

Kernel Engineer

Cerebras Systems · Software

US / Canada

Kernel-software headcount expanding across geographies at the same time — the same signal as the staff-level req, one rung down.

Negotiable

Director, Energy Market Development

CoreWeave · Energy

New York, NY

Director-level ownership of energy markets confirms this is a strategic function, not a single hire.

Negotiable

Staff Kernel Optimization Engineer

Cerebras Systems · Software

Remote

Kernel work for wafer-scale hardware opened fully remote — unusual for hardware-adjacent roles, which suggests Cerebras is widening its funnel for a scarce skill.

Negotiable