Summary

The AI Inference Engineer will own Fuse Energy's inference serving layer, building request routing, batching, scheduling, autoscaling, and model optimisation for high-throughput workloads. The role works closely with CUDA and GPU engineering teams, while the working arrangement is not specified.

Responsibilities

• Define inference serving strategy and architecture
• Design and build high-throughput serving stack
• Optimise models with quantisation and speculative decoding
• Select serving frameworks and orchestration approaches
• Translate performance commitments into capacity plans
• Integrate custom kernels with CUDA engineers
• Establish serving benchmarks and reliability standards