Company
Thunder Compute is GPU cloud platform built on custom virtualization that delivers 5x better margins than competitors. Think of us like VMWare for GPUs, monetized with a cloud platform.
4-person team, $4.5M raised, 100%+ MoM revenue growth. 6 days per week in-person in SF.
We are building a cloud platform backed by our virtualization stack, where we network-attach GPUs over TCP through API remoting. This lets us oversubscribe hardware aggressively; we have a pool of GPUs used by 3-4 times more servers than other platforms.
If building a ChatGPT wrapper or selling out to a hedge fund makes you shudder, there’s a third option.
Role
You'll work on the C++ core of our cloud platform: performance optimization, systems debugging, and research into oversubscription, checkpointing, and related problems.
You will spend your days in the weeds of complex, performance-critical systems in production.
Core technical skills:
- Top 0.1% C++ ability
- Experience optimizing NIC/GPU performance
- Deep understanding of low-level networking, compilers, or similar systems
- Experience in latency-sensitive environments
Must haves:
- Self-driven (you'll own your work from day 1)
- High speed-to-output
- Potential for senior leadership as we scale
- Strong communication skills
Preferred
- CS PhD/Masters from top 50 US university, 3.5+ GPA (Bachelor's required)
- 3+ years at trading firm, hardware company (NVIDIA, SambaNova), systems research or similar
Compensation
- Full-time, generous salary + equity.
- You'll report to Co-founder/CTO Brian (ex-Citadel quant dev).