Inference Engineer

The role

You will own the cost and performance of the inference stack. You will work on caching, batching, quantization, decoding, and kernel-level optimization to improve throughput and latency.

What you will do

• Improve throughput, cost, and tail latency through KV-cache management, continuous batching, speculative decoding, and quantization.
• Optimize long-context prefill and decode workloads based on real production traffic.
• Tune routing between infrastructure and external providers based on cost, capacity, and performance.
• Work within serving engines such as vLLM, SGLang, and TensorRT-LLM.
• Build profiling and measurement systems that show where time, memory, and compute are being spent.

Required qualifications

• 5+ years in ML systems, inference infrastructure, or performance engineering, with measurable improvements in cost or latency.
• Deep understanding of model serving, including prefill and decode, memory bandwidth, batching, and concurrency.
• Production experience with serving engines such as vLLM, SGLang, or TensorRT-LLM.
• Strong Python skills.
• Proficiency in C++, Rust, or another systems language.
• Experience with GPU performance, including CUDA, NCCL, mixed precision, memory layout, kernels, or quantization.

Benefits

• Innovative culture
• Growth paths
• Adaption Passport: An annual travel allowance to explore somewhere new
• Weekly allowance for takeout or grocery delivery
• High-quality medical coverage and benefits
• Team offsites and visits to SF HQ

About the company

adaption builds efficient intelligence that evolves in real-time. The team is small, talent-dense, and operates with high velocity and extreme ownership.