> Markdown version of [/jobs/ext/3550208-ai-infrastructure-engineer](https://www.wearedevelopers.com/jobs/ext/3550208-ai-infrastructure-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # AI Infrastructure Engineer - **Company:** Propio LS LLC - **Location:** Overland Park, KS, United States - **Experience:** Experienced - **Contract:** Permanent contract - **Skills:** Artificial Intelligence, Amazon Web Services, Amazon Elastic Compute Cloud, Amazon S3, Configuration Management, Disaster Recovery, Distributed Computing Environment, Distributed Systems, Identity and Access Management, Job Scheduling, Open Source Technology, Site Reliability Engineering Practices, Management of Software Versions, WebSocket, WebRTC, AI Infrastructure, Delivery Pipeline, Large Language Models, Amazon Virtual Private Cloud (VPC), AI Platforms, Low Latency, ONNX (Open Neural Network Exchange) Format, Live Streaming, Slurm, Machine Learning Operations, TensorRT, Hardware Infrastructure, Cloudwatch, Decoding, Grpc - **Published:** September 30, 2026 - **Apply:** https://www.indeed.com/viewjob?jk=6a97a6f591abd907 ## About the Role * 3+ years of experience building and operating AI/ML infrastructure, inference platforms, or distributed systems in production. * Hands-on production experience with AWS, particularly EKS, EC2 GPU workloads, ECR, S3, IAM/KMS, VPC networking, and CloudWatch and/or OpenTelemetry. * Hands-on experience with at least one inference stack, such as vLLM, SGLang, TensorRT-LLM, Triton, KServe, or Ray Serve. * Experience operating ML/LLM systems in production, including model serving, autoscaling, monitoring, incident response, and performance benchmarking. * Familiarity with GPU infrastructure and at least one serving stack, such as vLLM, SGLang, TensorRT-LLM, or Triton. * A working understanding of LLM Ops practices, including evaluation, observability and tracing, cost control, and versioning., * Low-latency, real-time, or streaming inference experience, especially for audio/speech-directly relevant to interpretation. * Experience with real-time audio pipelines and streaming protocols, including WebRTC, WebSocket, and gRPC streaming. * Experience supporting distributed training environments using FSDP, DeepSpeed, Megatron-LM, Slurm, EKS, or SageMaker HyperPod. * Inference optimization experience, including quantization, KV-cache optimization, speculative decoding, and continuous batching. * Experience building Edge AI deployment toolchains using ONNX Runtime, TensorRT/Jetson, ExecuTorch, llama.cpp, MLC-LLM, or similar runtimes. * Experience with SRE practices, capacity planning, production incident response, and secure infrastructure for PHI/PII. * Contributions to relevant open-source infrastructure, serving, observability, or edge-runtime projects. Even if your experience doesn't perfectly match every qualification, we encourage you to apply. We're looking for potential, drive, and a commitment to growth as much as experience. ## Description As an AI Infrastructure Engineer, you will primarily design, build, and operate our AWS-based, GPU-accelerated inference and streaming serving platform, optimizing it for low latency, high concurrency, reliability, and cost. You will also support the research team's training environment, further the AI team's ML/LLMOps capabilities, and enable the development of edge AI deployments. You'll be empowered to: * Take ownership of important initiatives and outcomes. * Drive meaningful business results. * Influence decisions and contribute new ideas. * Partner with talented, high-performing team members. * Challenge yourself through continuous learning and growth. * Help shape the future of a rapidly growing organization What You'll Own * Real-time inference and serving. Design, deploy, and operate low-latency streaming inference systems on AWS for LLM, ASR, TTS, and multimodal models. Optimize time to first token/audio, p95/p99 end-to-end latency, real-time factor, throughput, concurrency, GPU utilization, and cost per stream using technologies such as vLLM, SGLang, TensorRT-LLM, Triton, TensorRT. * Training environment. Support and evolve the research team's AWS-based training environment, including reproducible containers, GPU job scheduling, distributed job execution, checkpoint and resume capabilities, experiment tracking, model and data artifact access, and researcher self-service. Enable workloads using FSDP, DeepSpeed, Megatron-LM, Slurm, EKS, or SageMaker. * ML/LLMOps. Build model and artifact registries, lineage and versioning, automated evaluation gates, deployment pipelines, shadow and canary releases, rollback workflows, runtime and configuration management, and production observability * Edge AI infrastructure. Partner with researchers and device and embedded teams to build repeatable model optimization, packaging, validation, and deployment workflows for resource-constrained edge targets. Support model export and compilation, post-training quantization, runtime integration. Reliability and security. Own capacity planning, production readiness, incident response, disaster recovery, and the secure operation of the AI platform. * least-privilege IAM, KMS encryption, private networking, secrets management. What Makes Someone Successful Here The most successful people at Propio aren't necessarily the ones with the longest resumes. They're the people who: * Take ownership instead of waiting for direction. * Embrace challenges as opportunities to grow. * Continuously seek better ways of working. * Turn ideas into action. * Hold themselves and others accountable to high standards. * Are driven by making a measureable impact.