Forward Deployed Engineer - AI Inference

Location: San Francisco, onsite

Salary: Up to $250,000 base salary plus equity.

Company

A well-funded AI infrastructure startup building a high-performance inference cloud for open models.

What you will do

• Profile, benchmark, and trace customer inference workloads.
• Identify bottlenecks across models, serving engines, kernels, hardware, and production infrastructure.
• Prove performance using customer models and production traffic.
• Build and deploy the engineering needed to bring workloads into production.
• Own latency, throughput, reliability, and error-rate outcomes across three to six customer accounts.
• Convert recurring customer problems into improvements for the core platform.

Requirements

• Strong systems engineering background with the ability to work directly with technical customers.
• Understanding of LLM inference fundamentals, including prefill, decode, latency, throughput, and related trade-offs.
• Experience with profiling, tracing, benchmarking, and identifying production bottlenecks.
• Ability to use GPU programming technologies such as CUDA, HIP, or Triton.
• Ability to explain technical constraints clearly, defend engineering methodology, and adapt conclusions based on evidence.
• Comfort working autonomously across the stack and taking responsibility for production outcomes.

Nice to have

• Experience with LLM inference and model serving.
• Distributed systems and infrastructure experience.
• Performance engineering and optimization experience.
• Knowledge of quantization, speculative decoding, heterogeneous accelerators, cluster operations, reliability, and observability.

Culture & Benefits

• Highly autonomous and technically intensive environment.
• Small, talent-dense team with direct influence on customer performance, product direction, and company growth.
• Meaningful equity included with the compensation package.
• Engineers move across the stack and solve problems without waiting for tightly defined specifications.