ML Compiler Backend Engineer

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 bot

Etched is looking for exceptional ML compiler backend engineers to join our team and address the problem of building production-grade software to allow users to run existing transformer models on our hardware with better performance and latency. The ideal candidate is a veteran at shipping easy-to-use software for developers, and has experience with both using LLMs in production and in how they work underneath.

Responsibilities:

• Design, develop, and maintain our compiler stack.
• Collaborate closely with hardware teams to comprehend architecture, specifications, memory controllers, boot loader, external IP, requirements, and boot sequence. 
• Conduct debugging, testing, and validation of our compiler on target platforms, boot loader reliability, and verification on hardware simulation environment. 

Requirements:

• 5+ years of experience writing production-grade software.
• Strong grasp of computer architecture, data structures, system software, and machine learning fundamentals.
• Able to write production-grade code in C++ and in Python.
• Experience with modern compiler IRs, including at least one of (LLVM, MLIR, Relay)
• Experience with PyTorch

Desired qualifications:

• Work experience at a cloud provider, AI company, or LLM startup
• Experience implementing SIMD algorithms on vector processors
• Experience with open source ML compiler frameworks such as MLIR
• Experience with inference servers/model serving frameworks (such as Triton, TFServ)
• Experience with optimized libraries for transformer inference (TensorRT-LLM, vLLM, DeepSpeed, HF Inference Endpoints)

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.