Forward Deployed Inference Engineer

About the role

The Forward Deployed Inference Engineer will bridge the gap between customer model requirements and Tensordyne's high-performance AI inference hardware. This role involves optimizing model performance, managing customer deployments, and collaborating with internal engineering teams to refine the platform based on real-world usage. You will own the technical path from initial workload assessment to successful production rollout.

Responsibilities

• Turn customer workloads into fast, credible performance answers through profiling and benchmarking.
• Define relevant KPIs, compare against competitive baselines, and keep evaluation methodology current.
• Convert and bring up customer models on the Tensordyne stack, validate numerical quality, and identify performance bottlenecks.
• Work directly with customers and partners on technical PoCs, integration, deployment, and debugging.
• Translate requirements into measurable acceptance criteria for quality, latency, and throughput.
• Track profiling-to-hardware accuracy by comparing results with actual hardware deployments and flagging missing capabilities.
• Turn repeated customer-specific learnings into reusable tooling, documentation, and product improvements.

Required skills

• AI models
• Inference systems
• LLMs
• Python
• PyTorch
• Profiling
• Benchmarking
• Model optimization

Nice to have

• vLLM
• SGLang
• Quantization
• Distributed inference
• Kubernetes
• Rust

Qualifications

• Strong hands-on experience with AI models and inference systems
• Experience profiling, benchmarking, or optimizing model inference
• Experience working directly with customers or external technical partners
• Experience bringing models up on new accelerators or non-standard hardware

About the Company

Tensordyne is building a new class of AI inference system designed for high-performance, power-efficient deployment of the world's most demanding generative AI workloads. The platform integrates purpose-built silicon, new AI math, optimized scale-up networking, and memory architecture for large-scale AI inference.