Principal AI Governance Architect
Core
Translate security, privacy, compliance, and architecture requirements into executable platform controls for AI workloads while establishing patterns for knowledge enablement, evaluation, telemetry, and audit evidence.
Role type
Principal AI Governance Architect (Director level)
Builds
Executable platform controls, governed knowledge patterns, audit evidence systems, and operational reporting dashboards for AI applications.
Domain
AI Security, Governance, and Observability
Deliverable
production ML models | product features | dashboards & analysis | infrastructure
Required skills
Cloud security, platform security, governance engineering, security architecture, data engineering, observability, analytics engineering, ML evaluation, AI application monitoring, identity management, secrets management, logging, audit trails, data classification, least-privilege design, evidence capture, retrieval-augmented generation, embeddings, vector stores, metadata, indexing, citation, access control, knowledge-source quality, automation, software engineering, AI tool usage for system building, documentation, policy mapping, risk analysis, quality metrics, control design, dashboard development, evaluation design, control review.
Preferred skills
AI governance, model risk management, LLM application security, agent security, data protection for AI systems, Amazon Bedrock, AWS IAM, CloudTrail, CloudWatch, PrivateLink, KMS, VPC design, OpenSearch, vector databases, BI tools, observability platforms, LLM evaluation, prompt evaluation, retrieval evaluation, golden datasets, regression testing, AI quality frameworks, Temporal, workflow orchestration, enterprise knowledge systems, document repositories, metadata governance, search relevance, permission-aware retrieval, PHI, PII, client-confidential, regulated, sensitive-data environments.
Technologies
Amazon Bedrock, AWS IAM, CloudTrail, CloudWatch, PrivateLink, KMS, VPC, OpenSearch, vector databases, Temporal, BI tools, observability platforms.
Responsibilities
Translate security, privacy, compliance, and architecture requirements into executable controls for AI workloads; Define workload classification patterns and required controls for each class; Establish prompt, response, embedding, retrieval, logging, retention, redaction, and client data segregation patterns; Define audit evidence patterns for model access, data movement, retrieval, tool calls, approvals, exceptions, and operational events; Design identity, secrets, network, sandbox, logging, and approval-gate patterns for AI applications and agents; Build governed knowledge patterns for authoritative sources, ingestion, indexing, metadata, access control, freshness, citation, and retrieval evaluation; Help select the first Huron Knowledge domain, source, or integration pattern for MVP validation; Define and implement retrieval quality metrics, model evaluation patterns, regression checks, operational telemetry, dashboards, and quality reporting; Partner with infrastructure engineers to implement controls, evidence, and reporting through automation rather than manual processes; Help teams understand whether AI systems are producing useful, grounded, safe, auditable, and cost-effective outputs; Use AI tools hands-on to accelerate control design, policy mapping, knowledge analysis, evaluation design, dashboard development, documentation, and evidence review.
Seniority
Director, hands-on IC with strategy & mentorship
