Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

Navan

NewBe an early applicant

AI Architect

Posted Updated

Description

  • We are building AI capabilities that improve important human workflows across the company
  • As AI Architect, you will set the technical direction for how we design, evaluate, integrate, and operate AI systems
  • You will help teams choose the right approach for each problem—from conventional software and deterministic workflows to retrieval-augmented systems and agents—and ensure those systems meet clear standards for quality, reliability, safety, and value
  • This is a strategic, hands-on role
  • You will establish reusable architecture and evaluation practices, guide cross-functional teams, and prototype or validate key patterns where needed
  • You will partner with product and engineering leaders to turn high-value opportunities into systems that can be measured and operated in production
  • Set AI architecture direction: define reference architectures and design principles for AI services, agents, model integrations, data and context flows, tools, and human review. Make trade-offs among quality, autonomy, reliability, latency, cost, and complexity
  • Build the evaluation discipline: help teams define expected behavior and success measures; design representative datasets and evaluation methods; write and maintain evals; establish regression checks and quality thresholds; and use results to prioritize improvements
  • Guide production readiness: shape shared practices for tracing and observability, versioning, monitoring, incident response, privacy, security, and lifecycle management of AI systems
  • Enable teams to ship: create reusable patterns, reference implementations, and practical guidance. Coach engineers through architecture reviews and help teams move valuable work from prototype to production
  • Connect architecture to outcomes: partner with product, operations, and domain teams to identify workflow needs and measure whether AI improves completion, quality, time, or other agreed outcomes
  • Stay close to implementation: prototype high-risk or foundational designs, review critical implementations, and use production evidence to refine architecture
  • Track the evolving ecosystem: evaluate models, agent runtimes, orchestration approaches, and evaluation tooling against our requirements, and recommend adoption based on evidence

Benefits

  • Parental leave
  • Commuter benefits
  • Healthcare
  • Flexible Vacation
  • Health and Wellness
  • Fuel for Connection
  • Pet Friendly
  • IATAN (travel-related discounts)
  • 401k (retirement plan)
  • Connectivity Allowance
  • Learning & Development
  • Experience defining quality expectations and building evaluations for AI behavior; able to connect offline evaluations to production outcomes
  • A track record of setting technical direction across teams and influencing decisions through evidence and collaboration
  • Practical understanding of LLM applications, including retrieval and context construction, tool use, agent patterns, model selection, and failure handling
  • Strong experience designing and delivering production software systems, with substantial work on AI/ML-powered products or services
  • At least 3 years of experience in AI architecture with demonstrated organization-wide impact
  • Clear communication across engineering, product, and executive audiences, with a builder’s mindset and comfort prototyping and reviewing implementation
  • Sound judgment about where AI is appropriate and when a simpler deterministic approach is better
  • At least 8 years of experience in backend software development
  • Experience with production concerns such as observability, privacy and security, reliability, cost, latency, and safe human escalation
  • Establishing shared AI platforms, service patterns, or architecture standards across multiple teams
  • Creating evaluation datasets, test harnesses, regression suites, or quality dashboards for LLM or agent systems
  • Operating AI products with measurable user or business outcomes
  • Working with cloud AI platforms, model APIs, agent frameworks, tracing/evaluation tools, and data platforms. Adaptable judgment matters more than one vendor’s tools

Skills

Apply

See also

Architecture jobs by country — openings, pay and top skills →