AI Engineer

We are building an agentic AI platform for enterprise teams, delivered as a single-tenant install into each customer's cloud. Every architectural decision has to work for one customer or fifty, in AWS, Azure, or GCP, behind their identity provider, on their data.

We're hiring a Staff (or Senior Staff) Engineer to own the technical direction of the platform. You'll set the architecture, lead the hardest engineering work hands-on, and partner with the Forward Deployment team to make sure what we build actually deploys cleanly into customer environments. This is an individual contributor role with broad authority — you won't manage people, but you will set the technical bar for the engineering team and be the deciding voice on cross-cutting decisions.

Requirements for the candidate
Must-Have Qualifications
- Staff-level scope — 10+ years engineering experience, with at least 3 at the staff level or equivalent (principal IC, technical lead of a major system, founding engineer). Title matters less than scope demonstrated
- Production LLM systems — has built, shipped, and operated LLM-based applications at scale. Prompt engineering, RAG, tool-using agents, prompt-injection defense, evaluation pipelines. Has managed cost and latency, not just prototypes
- Multi-tenant SaaS architecture — has shipped a multi-tenant product and knows the trade-offs between shared-everything, shared-database, and silo deployments. Has opinions on when each is right
- Distributed systems foundation — idempotency, eventual consistency, queue-based async, distributed locking. Has implemented or operated these in production, not just read about them
- Security architecture — has designed an authn/authz system end-to-end. OWASP top 10 as patterns you've prevented in code reviews, not a list to memorize
- Deep proficiency in Python; comfortable enough in TypeScript/Node to read and review
- PostgreSQL under load — pool tuning, indexes, replication, online migrations. Working knowledge of pgvector or comparable vector store
- Cloud-native architecture — strong in AWS at minimum, with working understanding of at least one of Azure or GCP. Knows what's portable and what isn't
- Decisive on hard calls — comfortable being the deciding vote on architectural questions when there's no clean answer

Candidate responsibilities
Platform Architecture
Own the architecture of the AI platform end-to-end — agent orchestration, tool execution, memory, retrieval, observability, identity, authorization
Make and document the irreversible decisions: data model, async patterns, caching strategy, deployment topology
Lead the LangGraph-based agent framework: supervisor patterns, specialist agents, evaluation harnesses, regression suites
Author and review every significant ADR; keep the architecture coherent as the product and team grow

Multi-Tenant & Customer-Cloud Foundations
Design data isolation, identity, and authorization to work in single-tenant customer-cloud installs from day one — and to scale to a managed multi-tenant offering when the time comes
Own the integration with WorkOS (multi-tenant SSO broker) and Cerbos (centralized policy decision point) — schema, deployment topology, policy bundle workflow, customer customization boundary
Define the boundary between platform-owned code and customer-customizable layers (policies, prompts, tools, schemas, model choices)
Partner with the Forward Deployment team to make sure every architectural choice has a clean install story across AWS, Azure, and GCP

LLM & Agent Safety
Lead the design of guardrails and AI safety mechanisms — prompt-injection defense, output validation, PII handling, evaluation pipelines
Own the integration with our Athena guardrail router and the safety contracts agents must meet before they ship
Drive cost discipline on LLM workloads — caching strategy, model routing, token budgeting, latency targets
Make every agent decision auditable — observable, traceable, replayable, and defensible to a customer's security team

Hands-On Engineering
Personally lead the hardest work — distributed systems edge cases, performance investigations under load, security architecture, novel agent patterns
Write code and reviews, not just docs. The IC bar at staff level here is high
Pair with engineers on hard problems; raise the technical level of every review you touch
Be on point for production incidents that touch architecture; own the postmortem

Mentorship & Cross-Functional
Mentor senior engineers; be the person they go to when they want to do something technically ambitious
Be the engineering voice in customer technical conversations — CISO reviews, architecture deep-dives, incident escalations
Partner with the Senior Manager, Forward Deployment on customer-facing technical decisions where deployment shape and platform shape have to agree
Partner with product to translate roadmap into technically coherent quarterly commitments and an honest dependency map