16 hours ago
Remote, WorldwideMid Level
Responsibilities
- Implement complex computational algorithms on GPUs and CPUs with demanding latency and throughput requirements.
- Refactor existing solutions to improve scalability.
- Develop and debug high-performance GPU and CPU applications.
- Design custom CUDA kernels and work with third-party libraries.
- Profile and analyze application performance using Nsight Systems, Nsight Compute, and nvprof.
Requirements
- Commercial experience developing and debugging high-performance GPU and CPU applications with a strong focus on latency and throughput.
- Hands-on experience with third-party libraries and custom CUDA kernel design.
- Proficiency with Nsight Systems, Nsight Compute, and nvprof.
- Strong understanding of data structures, algorithms, and object-oriented programming in C++.
- Ability to work effectively in remote or hybrid teams with variable, project-based responsibilities.
- Curiosity and proactive engagement with emerging GPU, HPC, and ML trends.
- Preferred: expertise in low-latency and real-time development, including sub-microsecond latency.
- Preferred: experience with lock-free data structures and associated design patterns.
- Preferred: proficiency in C, Linux driver development, and kernel-level components.
- Preferred: strong understanding of Linux system internals and networking.
- Preferred: knowledge of GPU memory optimization techniques including shared memory, registers, coalesced access, and persistent kernels.
- Preferred: experience with inference optimization frameworks.
Benefits
- Remote work from anywhere in the world.
- Access to global offices at any time.
- Flexible schedule.
- 40 paid days off.
- Competitive salary.
- Welcoming international team and challenging work with fast feedback loops.
- Access to cutting-edge hardware and technology.
About Eqvilent
Eqvilent is a proprietary quantitative trading firm that provides liquidity across global financial markets. It designs and runs high-speed trading systems that apply deep learning, FPGA acceleration, and signal processing to price and execute trades, earning revenue from market making and other proprietary trading strategies. Founded in 2020 and headquartered in Dubai, the privately held company operates across multiple jurisdictions and maintains an international team.