Mayan Weiss Coreweave

• Operator role-type pattern → likely focused on execution and delivery over strategy decks; will want to talk specifics, not vision slides.
• Content themes include team building and hiring → probably managing a team and actively thinking about org design and engineering culture, not just individual-contributor work.
• Focus on inference systems optimization → likely cares deeply about performance benchmarks, latency, and cost-per-token metrics; detail-oriented on technical tradeoffs.
• Based in San Jose at a GPU cloud company → embedded in the hardware-proximate layer of AI infrastructure, closer to silicon constraints than application-layer concerns.
• Occasional public writing signal → not a heavy self-promoter; likely values substance over visibility, so cold outreach needs to earn attention quickly.

Openers

• CoreWeave's positioning as the high-density GPU cloud puts Mayan's team at the sharp end of inference optimization — open by referencing a specific challenge in serving latency-sensitive AI workloads at scale.
• Hiring and team building show up as active themes — if your offering touches engineering org tooling, onboarding, or technical recruiting, that's a live hook.
• GPU infrastructure efficiency (utilization, scheduling, cost-per-token) is a core concern at his layer — any capability that moves those numbers is worth surfacing early.

Discovery questions

1. What does the biggest bottleneck look like right now in your inference pipeline — is it hardware allocation, software scheduling, or something further up the stack?
2. As CoreWeave scales, how are you thinking about building the engineering team — are you hiring for deep GPU specialization or more generalist systems engineers?
3. Where do you see the inference optimization problem evolving over the next 12 months — more a hardware problem, a software problem, or a workload-shaping problem?