Forward Deployed Inference Engineer
Core
Own the path from customer workload to optimized model running on Tensordyne hardware, acting as a technical bridge between engineering and customers.
Role type
Forward Deployed Inference Engineer
Builds
Optimized inference deployments for LLMs, VLMs, and diffusion models on Tensordyne silicon
Domain
AI Inference Systems / Hardware Acceleration
Deliverable
production ML models
Required skills
AI model architecture (dense/MoE LLMs, VLM, speech, diffusion), Python, PyTorch, profiling & benchmarking, systems optimization, problem-solving
Preferred skills
LLM serving stacks (vLLM, SGLang), customer-facing technical work, new accelerator bring-up, production inference (quantization, distributed inference, Kubernetes), Rust
Technologies
PyTorch, vLLM, SGLang, Rust, Kubernetes
Responsibilities
Profile and benchmark customer workloads against competitive baselines; Convert and bring up customer models on the Tensordyne stack; Execute technical PoCs, integration, and debugging with customers; Track profiling-to-hardware accuracy and flag missing capabilities; Turn customer learnings into reusable tooling and documentation
