Senior GPU Engineer

ZK proofs are one of the most critical technologies to blockchain scaling, interoperability and privacy, but are too complex for most developers today. Succinct’s mission is to make zero knowledge proofs accessible to any developer.

SP1 is a state of the art zkVM with performance that rivals custom ZK circuits that lets developers use ZK with normal programming languages, like Rust. The Succinct Prover Network, under active development, is a hosted infrastructure layer for any application to outsource proof-generation for open-source proof systems.

Today, top teams in the space, including Celestia, Wormhole, Lido, Avail, Near, and Gnosis are using Succinct’s infrastructure to build with ZKPs. We have a track record of shipping excellent products and cutting-edge technical work in ZK and have significant revenue from our customers.

We recently announced our $55 million raise across our Seed and Series A led by Paradigm, the best investor in crypto. Our team is still small but incredibly strong, and we are looking for exceptional people who are excited to work hard on challenging and impactful problems in a fast-paced environment. Our office is located in San Francisco and we have a strong preference for candidates based there for this role, but are open to exceptional remote candidates.

We are seeking a Senior GPU Optimization Engineer to design, develop, and implement deep performance optimizations and GPU-accelerated software solutions that will dramatically enhance the efficiency and speed of our zkVM, SP1. You'll also play a pivotal role in advancing the capabilities of our hardware-accelerated prover network. In this role you will:

• Identify, analyze, and resolve GPU performance limitations through meticulous performance profiling and optimization.

• Conduct in-depth analysis to identify latency bottlenecks at both hardware and software levels.

• Develop and optimize low-level GPU kernels using CUDA, PTX, and other NVIDIA-specific technologies.

• Employ performance analysis tools such as NVIDIA Nsight Systems and Nsight Compute to continuously profile and enhance GPU workload performance.

• Research and integrate advanced GPU computing technologies and methodologies to maximize system performance.

• Collaborate closely with engineering teams to enable rigorous functional and performance testing and validation.

• Expert understanding of GPU and CPU architecture with deep expertise in performance optimization and low-level hardware interactions.

• Strong proficiency with GPU software stacks, compiler optimizations, CUDA, PTX assembly, and low-level APIs.

• Extensive hands-on experience using performance analysis tools including NVIDIA Nsight Systems and Nsight Compute.

• Proven track record in identifying and eliminating performance bottlenecks, minimizing latency, and enhancing GPU workload efficiency.

• Demonstrated ownership mentality with strong communication and team collaboration skills.

Nice to have

• Experience programming in Rust or C++

• Experience working on OS kernels and/or writing device drivers

Everyone on our team is located in San Francisco and works together in-person. Ideally, we’re looking for someone who enjoys in-person work and is already in San Francisco or willing to relocate (we can cover relocation costs). For senior candidates who are an exceptional fit, we’d consider a remote position as long as you’re available to onboard in-person for the first few months.

• Above-market salary and generous equity compensation

• Health, dental, and vision insurance for employees and their dependents

• Lunch and dinner provided at the office

• Optional, company-paid travel to events and conferences

Even if the listed qualifications don't seem like an exact match, passionate and enthusiastic people who love to quickly learn are always welcome! Feel free to reach out or apply regardless and we would love to chat.