Director of AI Engineering

Director of AI Engineering CampusEvolve | Remote (U.S.) | Full time | Applications close November 30Compensation: $200,000 - $275,000 base salary, depending on experience Our Director of AI Engineering builds Pathways, an AI-native guide that helps people take real next steps toward education, careers, and economic security, and builds the team's ability to ship it with AI. CampusEvolve (campusevolve.ai) is a Seattle-based, mission-driven AI startup backed by the Gates Foundation, Schultz Family Foundation, and GitLab Foundation. Pathways is one tool for a person's whole pathway: from high school into a two-year or four-year program, a certificate, an apprenticeship, or the military, and on into work. People most often fall off a pathway in the gaps between programs, so Pathways is built to carry them across those gaps. Its retrieval layer covers more than 6,000 post-secondary credential programs in Washington State, plus financial aid, career guidance, and basic-needs resources like housing, food, and transportation. The north star is economic mobility, and at CampusEvolve, the mission is the culture. This is a senior leadership role that's about 80% hands-on building. You deliver the product rather than make sure it gets delivered, and you make engineering decisions, not recommendations. You'll lead three full-time engineers, work closely with the Founder/CEO and CTO, and partner with a full-time product manager and the VP of Community, who runs a network of 100+ co-design partners across K-12, post-secondary, and workforce programs. State agency partners are ready to roll out Pathways once it's engaging and safe, and you'll lead the engineering that gets it there. What You'll Work OnAlongside the building, you keep engineering, product, partners, and leadership moving together, and you work directly with end users far more than most engineering leaders do. • Ship simple, mobile-first AI experiences that turn real learner and job seeker problems into completed next steps, with deep context awareness, more agency for users, and handoffs to human advisors when they need one • Build AI-native and multiply the team's output. Use coding agents yourself to ship working software fast, coach engineers in using them well, and help define the team's agentic development lifecycle: clear objectives and acceptance criteria, adversarial multi-model passes on the spec, and a strong evaluation harness • Connect product analytics, AI evaluation, synthetic-user tests, user research, and real task outcomes to fast release decisions, and show improvements with evidence • Create reusable conversation, planning, resource-grounding, state, and action patterns that can be configured quickly for new personas, scenarios, localities, and partners • Take features through controlled production, owning architecture, data and state integrity, test automation, observability, performance, cloud cost, and release quality • Architect identity and access for learners, advisors, and partner institutions, including privacy-preserving sign-in, consent, delegated access, and role-based authorization. Outside security and privacy specialists support hardening, reviews, and compliance, and you own integrating their work into the product • Be the technical voice with funders and state agency partners, including answering their hardest questions about safety and outcomes What Success Looks Like • Within the first months, Pathways reaches high school students after a few rapid cycles of engineering and hardening • Within six months, it serves thousands of high school users • Within twelve months, it's adapted for many personas, each tied to a real outcome such as enrolling in a post-secondary program, starting an apprenticeship, or getting a job, on the way to about 100,000 users in Washington State • The team ships faster and with higher quality because of how you build and how you coach What We're Looking ForWe place more weight on what you've built, your impact on users, your production judgment, and your ability to raise the level of your teammates than on executive titles, pedigree, or years managing large teams. No degree required. • ou've personally built and shipped LLM-enabled applications used by real people, as a founder, founding engineer, staff or principal engineer, or deeply hands-on engineering leader • Advanced fluency with multi-model agentic development tools (Claude Code, Codex, or similar), plus the ability to inspect, test, debug, and improve AI-generated code • Practical experience with LLM orchestration and tool calling, retrieval and grounding, persistent context, APIs, reliable application state, and full-stack cloud deployments • Strong product instincts. You can design for teenagers and adults with different levels of experience, find the friction, simplify, and move users from exploring to finishing a real step • The judgment to make an early product secure and scalable without overengineering it, from trust boundaries