Senior Kernel Developer

Etched is building the hardware for superintelligence.

GPUs and TPUs are flexible AI chips that can run many kinds of models: CNNs, RNNs, LSTMs, and more. But today, almost all AI workloads, from ChatGPT to self-driving cars, are done on one model architecture: transformers. Using flexible AI chips for transformers is very inefficient: <5% of the transistors on an H100 are used for matrix multiplication!

Etched is building a single-purpose chip exclusively for transformer inference. We only support transformers, but in exchange our chips have an order of magnitude more throughput and lower latency than an H100. With Etched, you can build products that would be impossible with GPUs, like tree-of-thought agents and ultra-low-latency audio chat bots.

We are seeking an experienced OpenCL/CUDA kernel developer with experience in embedded platforms to join our team. The ideal candidate will be responsible for developing and optimizing kernels for computing on specialized hardware platforms for parallel processing applications. 

Responsibilities:

• Design, developed, and optimized kernels for parallel processing for hardware acceleration.
• Collaborate with domain experts to translate algorithms into efficient parallel processing.
• Profile, analyze, and optimize kernel and overall performance to achieve maximum throughput and efficiency on acceleration hardware. 
• Implement and optimize data movement between CPU and various hardware blocks for effective utilization of parallel processing on an embedded platform. 
• Implement synchronization mechanisms for data coordination between CPU and hardware acceleration, ensuring effective execution.
• Develop and maintain test suites and benchmarks to validate kernel functionality, performance, and stability within embedded RTOS setup.

Requirements:

• Experience in developing OpenCL or CUDA kernels for parallel computing. 
• 5 years low level programming eg C/C in HPC or embedded environments.
• Proficiency in programming languages such as C/C and expertise in OpenCL or CUDA.
• Strong understanding of parallel programming, data structures, and optimization techniques suitable for embedded systems.
• Experience with performance profiling and optimization tools for embedded systems.
• Familiarity with memory management, data parallelism, and vectorization techniques.

Desired qualifications:

• Experience working in hardware simulation/emulation environments.
• Experience with implementing / serving LLMs 

Benefits:

• Competitive salary and equity package
• Full medical, dental, and vision packages, with 100% of premium covered
• Work with world-class people and state-of-the-art AIs everyday

Etched is committed to fair and equitable compensation practices. Compensation is determined based on your qualifications and experience. Compensation packages also include generous equity in Etched.

We are an equal opportunity employer and do not discriminate on the basis of race, religion, national origin, gender, sexual orientation, age, veteran status, disability or other legally protected statuses.