Virtualization software to make GPUs 5x cheaper
Thunder Compute is a Y-Combinator backed AI-infrastructure startup. We are creating technology to improve GPU utilization through sharing and oversubscription at the CUDA API layer.
Our core software lets us network-attach GPUs over TCP. This lets us allocate compute wherever it is needed most, leading to utilization gains of 5x+, with minimal performance overhead. We believe servers using dedicated GPUs is archaic, and plan to scale our virtualization software to every data center GPU in the world.
This is a $100B+ opportunity, with a problem that uniquely boils down to technical execution. In this systems engineering role, you would be employee #2, working on our core virtualization software.
If you are a top 0.1% systems engineer, looking for a problem that can be solved through sheer technical execution, this is the single highest potential opportunity on the market.
You will report directly to Co-founder and CTO Brian (ex-citadel quant dev). You likely have prior experience in quant dev, at NVIDIA, or similar. You are fed up working as a cog in a machine, you’d rather make a bet on your technical chops and change the world than take the $600K TC alternatives sitting in front of you.
If this role is a fit, you know who you are.
This roll is full-time, on-site at our headquarters in Atlanta, GA, salaried with stock options. Thunder GPU, Inc. is an equal opportunity employer, and adheres to all applicable employment laws.
Thunder Compute is a cloud computing platform that makes GPUs cheaper and easier to use. We do this with proprietary software to virtualize GPUs over a network.
Benefits: