Source description
About the role
Evaluate candidate Large Language Models (LLMs), embedding models, and AI inference solutions for quality, latency, memory utilization, reliability, and operational performance on embedded GPU-enabled edge compute platforms, including the X9 Spider Mission Computer architecture.
Tune and optimize AI model runtime configurations for edge deployment, including quantization strategies, batching configurations, context window sizing, cache behavior, inference scheduling, and GPU memory utilization specific to operational edge hardware environments.
Collaborate with customer stakeholders to assess mission requirements and evaluate alternative edge compute platforms when operational demands exceed X9 Spider capabilities or when cost, performance, power, size, weight, or thermal tradeoffs require additional analysis.
Benchmark agentic AI workflows, inference pipelines, and model-serving architectures against target hardware constraints and operational performance thresholds.
Recommend model-selection, runtime, and configuration tradeoffs balancing mission effectiveness, latency, throughput, resource utilization, reliability, and operational sustainability.
Build and maintain repeatable performance and stress-testing frameworks for evaluating latency, throughput, tool-call overhead, failover behavior, degraded-resource conditions, and disconnected operational scenarios on edge compute platforms.
Package, deploy, validate, and sustain local model-serving components and inference services to support reliable operation within tactical and edge environments.
Collaborate with agent engineers, AI developers, and integration teams to validate that agent behavior, workflow reliability, and operational outcomes remain acceptable following model compression, quantization, runtime optimization, or hardware configuration changes.
Support deployment, troubleshooting, optimization, and sustainment activities for AI-enabled applications operating in edge, airborne, tactical, or disconnected operational environments.
Train customer technical personnel on supported model profiles, operational constraints, runtime tuning considerations, deployment limitations, troubleshooting procedures, and platform sustainment best practices.
Maintain technical documentation, benchmarking results, model validation reports, deployment procedures, optimization baselines, configuration guides, and operational support materials.
Support DevSecOps and CI/CD activities associated with AI model packaging, deployment automation, runtime validation, and operational release processes.
More at NextGen Federal Systems