Performance Modeling Engineer, AI Inference Deployments
New
P
Positron AIAI inference hardware
Source API remote eligibility restrictions: United StatesFull-TimeSenior
Salary220,000 - 270,000 USD per year
Apply NowOpens the employer's application page
Job Details
- Experience
- 5+ years of experience in performance modeling, capacity planning, or systems architecture for large-scale AI inference, HPC, or data-center systems.
- Required Skills
- PythonC++Rust
Requirements
- Have 5+ years of experience in performance modeling, capacity planning, or systems architecture for large-scale AI inference, HPC, or data-center systems.
- Have demonstrated experience building analytical or queueing performance models and calibrating them against measured systems.
- Understand LLM inference at the systems level, including prefill versus decode, KV cache, parallelism strategies, mixture-of-experts routing, batching, time-to-first-token, and tokens per second.
- Have working knowledge of data-center networking, including latency, bandwidth, and topology, and GPU performance characteristics.
- Have strong Rust skills, or expert-level C/C++ skills and willingness to work in Rust day to day.
- Be proficient in Python for analysis, experiments, and visualization.
- Be able to work from incomplete information, state assumptions explicitly, and provide estimates with error bars.
- Explain sizing results, their drivers, and their confidence to engineers and customers.
- Hands-on experience sizing or benchmarking production inference deployments across many accelerators is preferred.
- A track record of building customer-facing performance or TCO tools is preferred.
Responsibilities
- Partner with customers and product to turn deployment questions into focused studies with clear assumptions, options, metrics, and recommendations.
- Compare the same deployment on competitor hardware and explain the performance differences.
- Design and deliver a standalone performance-estimation tool customers can run on their own workloads.
- Model large-scale inference deployment topologies, including parallelism strategies, batching, context length, and mixed Positron-plus-GPU configurations.
- Model interconnect latency and bandwidth, redundancy, and system-level effects on throughput as chip count grows.
- Develop analytical and simulation-based deployment models that use the chip performance simulator.
- Validate projections against measured hardware runs and calibrate models as chips, software, and workloads evolve.
View Full Description & ApplyYou'll be redirected to the employer's site