> Markdown version of [/jobs/ext/1333014-platform-support-architect](https://www.wearedevelopers.com/jobs/ext/1333014-platform-support-architect). 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). --- # Platform Support Architect - **Company:** DDN, LLC - **Location:** Sacramento, CA, United States - **Experience:** Expert - **Contract:** Temporary contract - **Skills:** Artificial Intelligence, Nvidia CUDA, Continuous Integration, Data Infrastructure, Data Security, Software Debugging, Linux, Elasticsearch, Ethernet, General Parallel File Systems, InfiniBand, Remote Direct Memory Access, Reliability Engineering, Prometheus, Runbook, Search Technologies, Toolchain, Ceph (Software), Graphics Processing Unit (GPU), Load Balancing, Large Language Models, Grafana, Prompt Engineering, Generative AI, HybridCloud, AI Platforms, Kubernetes, Deployment Automation, Machine Learning Operations, TensorRT, Virtual Agents, Nim (Programming Language), Docker, Microservices - **Published:** July 18, 2026 - **Apply:** https://www.indeed.com/viewjob?jk=7bc1680d898037bd ## About the Role * 5+ years in Linux-based infrastructure roles (SRE, MLOps, platform engineering, or L2/L3 support) supporting production systems; 8+ years total technical experience preferred. * Strong hands-on experience with containers and Kubernetes (Docker/containerd, Helm, Operators; debugging pods, DaemonSets, CSI, CNI, and ingress/load balancers). * Demonstrated experience operating GPU-accelerated workloads in production: + NVIDIA GPUs, drivers, CUDA concepts, GPU utilization/perf triage + NVIDIA GPU Operator and Kubernetes-based GPU lifecycle management + Familiarity with DGX / HGX or similar GPU cluster platforms. * Practical experience with AI storage and networking for HPC/AI clusters: + High-performance storage systems (e.g., EXAScaler/Lustre, GPFS, Ceph, distributed object storage, enterprise NAS/SAN). + RDMA-accelerated and/or high-speed Ethernet/InfiniBand networking, including fabrics, switch topologies, and large-scale deployments. + Hybrid cloud or cloud-adjacent patterns (Kubernetes CSI, cloud-native fabrics, data locality). * Experience with one or more vector databases (Milvus, Qdrant, Pinecone, pgVector, OpenSearch/Elasticsearch vectors, etc.), including schema design, ingestion, and operations. * Solid understanding of RAG and Generative AI workflows: embeddings, retrieval, reranking, prompt design, context management, and how these interplay with vector search and GPU inference at scale. * Familiarity with NVIDIA AI Enterprise components and toolchain, for example: + NVIDIA NIM inference microservices + NVIDIA NeMo framework / NeMo Retriever / NeMo Curator + Triton Inference Server, TensorRT / TensorRT-LLM, CUDA libraries + NVIDIA blueprints for enterprise RAG and agentic AI. * Experience designing, operating, or supporting MLOps / GenAI pipelines: CI/CD for models, deployment strategies, canarying/rollback, GPU resource management, monitoring and alerting for AI services. * Strong diagnostic skills across Linux, containers, Kubernetes, GPUs, storage, and networking; able to quickly narrow fault domains and propose experiments or configuration changes. Support, architecture, and stakeholder skills * Track record of building reusable technical assets (runbooks, KBs, implementation guides, benchmarks, PoC templates) that improve support readiness and partner/customer success. * Excellent communication skills, capable of clearly explaining complex AI platform topics to both engineers and executive stakeholders, internally and with partners., * Prior experience with scale-out storage in GPU/AI environments. * Direct experience crafting and operating RDMA-accelerated HPC/AI clusters at scale, including spine-leaf or fat-tree network designs and large switch/router deployments. * Hands-on work with NVIDIA reference blueprints (Enterprise RAG, VSS, AIQ, industry-specific blueprints) or similar enterprise AI architectures. * Familiarity with AI observability and responsible AI practices (guardrails, monitoring for drift/toxicity, basic understanding of regulatory considerations like GDPR/HIPAA in the context of AI systems). * Experience with observability stacks (Prometheus, Grafana, Loki/ELK, NetQ, etc.) tuned for AI workloads, including service-level dashboards and SLOs. ## Description * Act as the primary NVIDIA AI Enterprise and vector database solutions expert for HyperPOD customer environments, bringing deep knowledge of NVAIE services (e.g., NIMs, NeMo, Triton, TensorRT/TensorRT-LLM, GPU Operator, licensing/NLS) and vector databases (e.g., Milvus) to guide diagnosis, optimization, and solution design. * Own complex end-to-end triage across GPU, NVAIE services, vector DB, Kubernetes, Docker, high-speed networking, and Infinia storage, distinguishing product defects from environmental and integration issues. * Diagnose and resolve performance bottlenecks in RAG and agentic AI workflows, from model selection and prompt/RAG configuration throughto vector search, GPU utilization, and data access patterns. * Collect and interpret logs and telemetry across Linux, containers, Kubernetes, GPU stack, vector DB, and storage/networking; build minimal repros and high-quality defect reports for escalation to NVIDIA, vector-DB vendors, OEMs, and internal engineering. Runbooks, diagnostics, and supportability * Author and maintain support triage runbooks and checklists for HyperPOD covering NVAIE services, Milvus/vector DB, GPU stack, Docker, Kubernetes resources, and their interaction with Infinia and the network fabric. * Define and validate unified diagnostics bundles that capture the right logs/configs/metrics from all relevant layers (Infinia, GPUs, NVAIE, Milvus, Kubernetes, network) to enable fast problem isolation and high-signal escalations. * Collaborate with observability and tools teams to shape Prometheus/Grafana/ELK/NetQ or equivalent dashboards that surface both platform health and RAG/service-level metrics (e.g., TTFT, retrieval latency, error rates, throughput). Enablement, PoCs, and reusable assets * Build hands-on labs and PoCs that mirror customer RAG and agentic AI use cases on HyperPOD, validating supportability and capturing "known good" configurations and troubleshooting patterns. * Develop reusable technical assets - implementation guides, best-practice playbooks, tuning checklists, example architectures - to accelerate time-to-value for customers, PS, and Support. Design feedback, readiness, and cross-functional leadership * Provide structured feedback from early field cases and PoCs into Product Management and Engineering on stack compatibility, upgrade order, rollback constraints, and observability needs for NVAIE, Milvus/cuVS, Infinia, and networking. * Collaborate closely with NVIDIA solutions architects, OEM architects, PS, and Support Innovation to align reference architectures and best practices with real-world support experience., Within 6-12 months, a successful AI Data Platform Solutions Architect will have: * Become the go-to internal expert for "how this AI and networking stack actually works in production" across Support, PS, Product, and NPI for HyperPOD. * Drive speed and quality of support at solution level; NVAIE, vector DB, and AI-workflow issues through high-quality diagnostics, architecture insight, and well-defined "golden stack" patterns. * Established clear, repeatable triage and escalation patterns for AI-side incidents that L1/L2 storage engineers can follow with confidence. ## Related Videos - [Docker network without Docker](https://www.wearedevelopers.com/videos/1418-docker-network-without-docker) - [Docker Compose: Rediscovered](https://www.wearedevelopers.com/videos/1978-docker-compose-rediscovered) - [5 steps for running a Kubernetes environment at scale](https://www.wearedevelopers.com/videos/88-5-steps-for-running-a-kubernetes-environment-at-scale) - [LLMOps-driven fine-tuning, evaluation, and inference with NVIDIA NIM & NeMo Microservices](https://www.wearedevelopers.com/videos/1582-llmops-driven-fine-tuning-evaluation-and-inference-with-nvidia-nim-nemo-microservices) - [Docker exec without Docker](https://www.wearedevelopers.com/videos/1094-docker-exec-without-docker) - [Docker build without Docker](https://www.wearedevelopers.com/videos/100114-docker-build-without-docker) ## Related Articles - [Graph and AI Trends 2026: Why Is AI Running but Not Yet Delivering?](https://www.wearedevelopers.com/magazine/680-graph-and-ai-trends-2026-why-is-ai-running-but-not-yet-delivering) - [Best US AI Conferences for CTOs in 2026: Build vs. Buy, Vendor Evaluation, and Peer Intelligence](https://www.wearedevelopers.com/magazine/736-best-us-ai-conferences-for-ctos-in-2026-build-vs-buy-vendor-evaluation-and-peer-intelligence) - [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it) - [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere) - [Got AI ideas but no money? Here are 10 free ways to level up your AI skills with Google Cloud](https://www.wearedevelopers.com/magazine/600-got-ai-ideas-but-no-money-here-are-10-free-ways-to-level-up-your-ai-skills-with-google-cloud) - [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)