Principal Engineer I - App Engg & AI

You will be designing the frameworks that lets a 700-person consulting firm ship governed autonomous systems repeatably across regulated clients. Responsibilities Own architecture decision records and reference architecture for the platform; drive review with enterprise architecture teams at Fortune 500 clients. Build production code. This role is roughly 70% hands-on — you set direction by shipping the hard parts, Define the evaluation strategy: offline eval harnesses, regression suites for non-deterministic systems, and the quality bar that gates promotion. Design multi-tenant isolation, identity, and authorization patterns (OIDC/PKCE, M2M, JIT provisioning) suitable for data-sensitive clients. App Development: Build functional applications using large language models (LLMs) and APIs (e.g., OpenAI, Anthropic, open-source Llama). Model and control unit economics — per-session token cost, agentic loop multipliers, model tiering — so engagements stay margin-positive. Mentor senior and lead engineers; raise the ceiling of the org, not just the output of a project. Partner with go-to-market on technical positioning, competitive differentiation, and executive-level architecture narratives. Required Qualifications 10+ years building and operating production distributed systems; 3+ years shipping LLM-based systems. Deep experience with at least one major cloud AI stack (AWS Bedrock / AgentCore, Azure AI Foundry, or equivalent) and fluency in the tradeoffs of the others. Strong grounding in one server-side language at production depth (Java/Spring Boot, Python, TypeScript) plus infrastructure-as-code. Practical experience with agentic patterns: tool use, multi-step planning, orchestration, human-in-the-loop approval flows, MCP or comparable protocol work. Track record designing for compliance-sensitive environments — audit logging, data residency, least-privilege authorization, policy enforcement. Ability to hold a technical argument with a client's chief architect and a business argument with their CFO in the same meeting. Preferred Prior work on compilers, type systems, policy engines, or CI/CD platforms — the mental model transfers directly. Consulting or professional services background; comfort with multi-client context switching. Experience with Snowflake, databricks or data governance tooling, or enterprise data quality platforms.