Agentic Operations Engineer

About Us

Integrate is building the first real-time, cross-boundary project management platform for the world’s most complex programs. We started with one of the most complex customers on earth, the U.S. government. Its programs span thousands of people, multiple organizations, billion-dollar budgets, and differentiated security environments. And somehow, too much of that work still runs on stale schedules, spreadsheets, hand-built slides, and project management software stuck in the Tetris era.

We’re changing that. Based in Seattle, Integrate has secured a $25 million U.S. Space Force contract and expanded across government and commercial programs, raised a $17 million Series A led by FPV Ventures, secured backing from J.P. Morgan, and built software that operates across government and commercial environments, from unclassified work to Top Secret networks.

We move fast, keep the egos low, and make room for quirks, strong opinions, fun, and the occasional chaos of building things from scratch with extreme velocity. We offer strong benefits, real ownership, and ambitious work with people you actually want to be around.

Our Solution

Integrate gives complex programs something they have never had before. One real-time schedule that can operate across organizations, security boundaries, and classification levels without fragmenting the schedule.

The program stays live as it changes. Teams can see how the program has moved, understand how a change ripples downstream, surface risk before it compounds, and control access down to the individual element. And our AI architecture works from that same continuously updated program state. It understands what is happening now, not a report assembled days or weeks ago. The schedule stops being a static plan and becomes a living reflection of the program itself.

We’re nowhere near done. We want to set a new standard for how the world’s most complex work gets done, and we won’t stop until we do.

About the Role

We're hiring an Agentic Operations Engineer to help us build the internal intelligence operating system of Integrate.

This role sits at the intersection of engineering, product, and operations. Your prime directive is to leverage modern AI tooling and agentic strategies to increase organizational velocity across engineering, product, design, and ops.

You are an engineer first. You've shipped real full-stack applications, and you understand the discipline of deploying and maintaining software people actually depend on. But what energizes you most is the next layer up: how do we structure a codebase, an organization, and a workflow so that humans and agents can collaborate fluidly? How do we eliminate the scavenger hunt for context and keep that context fresh? This is your purview.

This role is part tactical and part strategic. Tactically, you'll write code, build internal tools, automate pipelines and workflows, and audit repos so that the agents working alongside us route correctly and produce high-quality output. Strategically, you'll look ahead, parse requests coming in from product, engineering, design, and ops, and merge them into efforts that make sure we aren't doing redundant or parallel work across teams. You'll audit teams, unblock bottlenecks, stand up new systems, and ship tools that compound across the org.

You'll report directly to the VP of Product and work closely with engineering, product, and operations. This role is in person at our Ballard office in Seattle.

What You’ll Work On

Build the internal "intelligence OS" of Integrate — the tools, agents, automations, and conventions that compound the team's productivity over time

Organize our repos, docs, and conventions so that AI coding agents route correctly, find the right context, and produce high-quality output

Stand up and maintain internal tooling — chatbots employees use to find answers, CLI tools that support day-to-day workflows, small web apps that smooth over operational friction

Build and own automations around the PR review process — code quality checks, agent-driven first-pass review, security and convention enforcement

Partner with engineering to identify code-quality issues, ambiguous structures, or stale conventions that slow down both humans and agents, and propose fixes

Support the product team as they get closer to the codebase, and the design team as they work more directly in feature branches — pairing, building scaffolding, refining prompting workflows

Evaluate, deploy, and govern bleeding-edge AI tooling: IDE agents, evals, model gateways, MCP servers, on-prem systems

Audit teams proactively. Surface friction. Propose system-level fixes rather than one-off patches

Be the person teammates come to with questions, ideas, and half-built experiments — and the person who leads when no one is asking

Plan three to six months ahead for what an AI-native Integrate looks like, including on-prem and locally-run agent infrastructure for our most security-sensitive workflows

Contribute occasionally directly to the core application codebase when needed

Tech Stack

Our core application is built on React, Golang, GraphQL, REST, PostgreSQL, and GitHub Actions, with Tailwind CSS on the frontend. Familiarity with any of these is a plus, but you don't need to know our stack to be successful in this role. You'll spend most of your time in the layer above — building tooling, agents, and automations that work alongside our engineers, not necessarily inside our product.

Experience

We're looking for someone with deep fluency in AI tooling and agentic systems, and enough full-stack engineering experience to build real things people depend on. You don't need to check every box — but here's the kind of background that sets you up well:

AI & Agentic Systems

Prompt engineering & LLM integration — Knows how to write effective system prompts, structure context windows, and integrate with LLM APIs (Anthropic, OpenAI, etc.). Not just "has used ChatGPT" — understands token limits, tool use, and how model behavior changes with prompt design.

Agent orchestration — Experience building multi-step agentic workflows using tools like Claude Code, LangChain, or similar frameworks. Understands how to chain tools, handle failures gracefully, and design loops that don't go off the rails.

Agentic security posture — Understands the specific attack surface that comes with giving agents tool access: prompt injection, over-permissioned tool scopes, secrets leaking into context windows, and unintended data exfiltration through third-party API calls. Knows how to scope what an agent can see and do, apply least-privilege to MCP server access, and design systems where a compromised or misbehaving agent can't do catastrophic damage.

MCP / tool & API integration — Comfortable wiring up internal and third-party MCP servers, REST/GraphQL APIs, and webhooks so agents can actually do things. The glue layer between LLMs and the rest of the world.

Knowledge base design & RAG — Can architect a retrieval system — chunking strategy, embedding models, vector stores, metadata design — so that agents and internal tools can query organizational knowledge usefully rather than just dumping docs into a pile.

Evaluations, metrics & observability — Can define what "working well" looks like before a tool ships and verify it stays that way in production. Writes evals and test sets for LLM behavior, instruments tools with usage metrics and logging, and builds enough observability that regressions surface before the team notices. Critically, knows when this rigor is actually warranted — not every internal tool needs a full eval harness, and good judgment about a tool's blast radius matters as much as knowing how to build one.

Engineering & Tooling

Lightweight full-stack / mini-app development — Can spin up a small web app or internal tool quickly. Doesn't need to be a frontend wizard, but should be comfortable enough to deploy something real that people actually use.