Inference & Workload Optimization Lead – Horizon AI

STN Incorporated

. Profile customer training and inference workloads to identify bottlenecks in data loading, collective communication, memory bandwidth, kernel efficiency, batching, and parallelism layout .

Posted 10/21/2026full-timeRemote • United StatesSeniorWebsite

Core Competencies

Role fit

Core Competencies

Use this summary to align your resume positioning with the role.

Demonstrates expertise in GPU performance engineering, including profiling, memory bandwidth, and kernel efficiency, while effectively tuning workloads for optimal performance across customer clusters. Proficient in utilizing advanced serving stack technologies and profiling tools to enhance model serving and inference capabilities.

Highest-signal resume keywords

Deep GPU Performance EngineeringServing Stack Depth with vLLM, SGLang, TensorRT-LLMDistributed Training and Inference ExperienceTuning Collective Communication with NCCLFluency with Profiling Tools like Nsight and DCGM

ATS Keywords

Tailor your resume

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills

GPU Performance ProfilingMemory Bandwidth OptimizationKernel-Level UnderstandingBatching and KV Cache ManagementQuantization Trade-OffsTensor, Pipeline, and Data ParallelismCUDA Kernel WorkBenchmark Design

Soft Skills

Customer CommunicationPerformance Explanation

Tools & Technologies

NsightDCGMNCCLVLLMSGLangTensorRT-LLM

Industry Keywords

Model ServingInference WorkloadsCollective CommunicationCost-Aware RoutingLatency-Aware Routing

About the role

Key responsibilities & impact

Profile customer training and inference workloads to identify bottlenecks in data loading, collective communication, memory bandwidth, kernel efficiency, batching, and parallelism layout
Establish performance baselines and define good performance for each hardware generation, starting with current-generation NVIDIA hardware
Convert optimization expertise into agent-executable fixes applicable across customer clusters
Work with committed-capacity customers, directly or through operating teams, on workload tuning
Own internal model serving as a secondary responsibility, including cost- and latency-aware routing and model version migration
Advise on GPU allocation between internal inference and customer workloads

Requirements

What you’ll need

Deep GPU performance engineering, including profiling, memory and bandwidth behaviour, and kernel-level understanding
Serving stack depth with technologies such as vLLM, SGLang, or TensorRT-LLM, including batching, KV cache management, and quantization trade-offs
Distributed training and inference experience at scale, including tensor, pipeline, and data parallelism
Experience tuning collective communication such as NCCL
Fluency with profiling tools such as Nsight, DCGM, and framework-level profilers
Ability to explain performance findings to customers and encode them for an agent
Nice to have: CUDA kernel work, open-weight model deployment at scale, benchmark design, and prior customer-facing technical experience