Amazon Sagemaker Jobs in Colorado (NOW HIRING)

Why This Role Matters
Open-source intelligence is now a primary feed for mission decisions, not a supplement. The volume of publicly available information has outrun the ability of analysts to exploit it by hand. Mission teams need AI systems that ingest, enrich, and reason over that data, and they need those systems to run where the mission runs.

That includes the hard places. Disconnected enclaves. Degraded networks. Air-gapped regions with no route to the public internet. The architecture that works in commercial AWS does not survive contact with a denied environment. Someone has to design the version that does.

Bluestaq builds on AWS across every classification level, from GovCloud to the Secret and Top Secret Regions. This role sets how we use that platform for AI-enabled OSINT. You decide how models, data, and pipelines run on AWS, connected or not. You own the structural decisions this capability is built on, from day one.

The Role
As the Principal Solution Architect for Bluestaq's OSINT mission capability, you design the AWS systems that turn open-source data into mission intelligence. You work at the intersection of AWS cloud architecture, applied AI, and intelligence tradecraft. You translate analyst requirements into AWS service selection, model design, data pipelines, and deployment patterns that hold up in defense environments.

This is an ownership role, not a coordination role. You are the technical authority for how these systems are built on AWS and where they run, and you are building that capability from the ground up. When the architecture holds under load and across classification levels, you are the reason. When it drifts, you are the first to flag it and the first to arrive with a plan.

What You Will Own

• Set the AWS reference architecture for AI and ML systems that ingest, enrich, and exploit open-source and publicly available information at scale.
• Architect around the service gaps and parity differences between commercial AWS and the government and classified regions.
• Build the AI stack on Amazon Bedrock and Amazon SageMaker, including model access, Bedrock Guardrails, Knowledge Bases, and managed endpoints.
• Design retrieval pipelines using Amazon OpenSearch, S3 data lakes for governed, least-privilege access.
• Orchestrate workflows with AWS Step Functions and AWS Lambda, and define CI/CD with infrastructure as code using AWS CDK and Terraform.
• Extend services and capabilities to the edge for disconnected and DDIL operations and local model serving.
• Define the deployment pattern for fully air-gapped enclaves, including mirrored registries, signed model imports, and local vector stores.
• Define data movement patterns and cross-domain solutions across classification levels and coalition partners.
• Apply OSINT tradecraft and operational security discipline to collection system design.
• Own model governance for air-gapped and classified environments, including provenance, integrity verification, and update processes.
• Partner with security and accreditation authorities to drive ATO and continuous ATO.
• Partner with engineering directors and program leadership to align this architecture with Bluestaq's broader AI platform strategy, so it becomes a pattern other programs can build on.
• Establish the practices and documentation that let Bluestaq scale this capability and bring engineers on as the mission grows.

Required Qualifications

• 10+ years designing and deploying production systems, including 5+ years architecting on AWS.
• Deep AWS experience across compute, networking, IAM, storage, and data services, including AWS GovCloud.
• Hands-on experience with Amazon Bedrock and Amazon SageMaker for production AI and ML workloads.
• Experience with AWS networking and security, including VPC design, PrivateLink, Transit Gateway, IAM Identity Center, and AWS KMS.
• Experience with infrastructure as code on AWS using AWS CDK, CloudFormation, or Terraform, plus CI/CD in restricted environments.
• Experience deploying AWS workloads in disconnected, degraded, or air-gapped (DDIL) conditions, including AWS Snowball Edge or AWS Outposts.
• Working knowledge of the differences between commercial AWS and the GovCloud, Secret, and Top Secret Regions.
• Familiarity with vector search, retrieval-augmented generation, and embedding pipelines on AWS.
• Working knowledge of OSINT and publicly available information collection, tradecraft, and analytic workflows.
• Working knowledge of DoD compliance frameworks, including the Cloud Computing SRG, FedRAMP, and ITAR.
• Current AWS Professional certification, such as Solutions Architect Professional.
• Clear technical writing and the ability to brief senior leaders and decision-makers.
• Prior OSINT analyst or PAI targeting experience in a DoD or Intelligence Community context.
• Completion of a foundational intelligence analyst course, such as the DoD Intelligence Analyst Course, the Army 35F course, or ODNI Analysis 101, or a CDASA certification.

Preferred Qualifications

• AWS Machine Learning Specialty or Advanced Networking Specialty certification.
• Experience with cross-domain solutions and multi-level security.
• Strong grounding in containerization and orchestration with Kubernetes and Docker on AWS, such as EKS.
• Experience with self-hosted model serving on local GPUs, such as vLLM or TGI, inside an enclave.
• Familiarity with the NIST AI Risk Management Framework and MITRE ATT&CK.
• Experience supporting OCONUS or deployed DoD programs.

Required Education

Bachelor's degree in Computer Science, Engineering, Data Science, or a related field, plus 10+ years of relevant experience. Equivalent experience may substitute for the degree, with 4 additional years beyond the minimum required in lieu of a Bachelor's degree.

Required Location & Travel
This position remote, open to candidates based in Colorado or Virginia, with up to 12 trips per year to customer, partner, or government sites, including secure facilities.