Why we exist
Getting permission to build is far harder than it should be.
Many cities still run permitting through PDFs, disconnected systems, tribal knowledge, and long email chains. Applicants struggle to understand what is required. Reviewers spend too much time chasing missing information. Projects stall.
The result is frustrated people, overwhelmed staff, and fewer homes and community projects getting built.
Govstream.ai exists to change that. We are building one continuous, conversation-driven workflow that guides a project from its first zoning question through application, review, and inspection.
Our north star is same-day construction permits, down from the many months a review can take today.
What we're building
Strip permitting to its essence and you find three things tangled together: a long-running conversation between the builder and the city, a workflow moving on both sides of the counter, and hundreds of decisions that need context and support.
Most permitting systems digitized the transactions and stopped there.
Govstream.ai rebuilt those three strands as one system. We call it Conversational Workflow for Permitting.
Govstream.ai is the only permitting system that brings all three together:
- It guides the builder at every step and carries the full context of the project—conversations, emails, documents, and permit history—across the life of an application.
- It gives city staff decision support across every role and discipline involved in review, without replacing their judgment.
- It manages communication between the city and the builder: translating code language into clear, practical guidance and bringing the builder's information back to staff in a form they can act on.
Underneath, that conversation produces a living set of facts about the project: what is being built, where, how tall, on what kind of lot, and what rules apply.
AI agents help draw those facts from conversations, plan sets, permit records, and GIS data. Every proposed fact is checked before it becomes part of the record, with full provenance of where it came from.
The city's rules live in House, our domain-specific language. As new information arrives, the system evaluates those rules consistently and predictably: What is now required? What has been satisfied? What is still missing?
The answer drives the next question for the applicant, the next action by the system, or the next item on a reviewer's screen.
Applicants experience this as the Permit Center: one guided conversation across web and email that helps them understand what is required and reach a complete, review-ready application.
City staff experience it as the Review Center: decision support that shows what is complete, what rule was triggered, and why, with the relevant code behind the answer.
One boundary holds everywhere: Govstream.ai determines what is required, never what is approved. Judgment stays with city staff. Our job is to give them better information and more time to apply it.
Every city has different codes, data, and working practices. We configure Govstream.ai using the city's own materials, then test the system against real historical permits and the decisions reviewers actually made.
Agents help draft and test the work, but people review it, resolve ambiguity, and own the result. When we find a miss, we fix it and turn it into a permanent test.
That is how the system becomes more accurate, and how each new city becomes faster to launch than the one before it.
Running that cycle with real cities, and improving the system as you go, is at the heart of this role.
The role
Every city begins with a signed contract. The work matters when residents and builders can use the service, city staff trust it, and city leaders can point to real improvements.
As the first Customer Success Engineer at Govstream.ai, you will own that journey from kickoff through configuration, validation, pilot, public launch, and expansion.
You will be the city's primary partner at Govstream.ai and the person accountable for keeping the work accurate, visible, and moving.
Launch is not the finish line. You will be a fierce advocate for customer outcomes: whether permitting is actually becoming faster and simpler, not just whether the software went live.
You will also advocate for the people using the product. You will listen closely, remove friction, and push for an experience that helps city staff and applicants feel informed, capable, and in control.
This is not a traditional customer success role. You might spend one hour debugging a rule or SQL query and the next helping city reviewers reach a decision about how their process should work. You need to enjoy both.
Today, our founders and engineers take cities live together. We have reached the point where this work needs a dedicated owner: someone who brings focus to every launch, keeps communication with cities clear and consistent, and builds the methodology and tools that future Customer Success Engineers will use.
The pace of Govstream.ai's growth depends on this work. A signed city creates little value until people can use what we have built.
Launch cities and earn their trust
You will model each city's permitting process in its own language, validate it with staff, run weekly working sessions, and keep every open question, commitment, and risk visible.
You will make launch-readiness calls, including telling a city that wants to go public that the system is not ready yet.
Every engagement builds toward evidence. You will test the city's configuration against its own historical permits and measure how many of the issues reviewers once had to chase could now be caught earlier.
In our backtesting, first-submission completeness increased from 22% to 67% in one city and from 39% to 70% in another.
From there, you will guide the city through a test environment, invited real applications, public launch, and expansion across more permit types.
Do the technical work
Most of an onboarding runs through Claude Code, our MCP tools, and automation we have already built.
Agents pull and reshape permit history, draft House configuration, and run test suites. You direct the work, inspect the output, and handle the parts that require human judgment.
That means reading the SQL and rules the agents wrote, checking what they got wrong, debugging from logs and traces, and distinguishing between a flawed rule and a data source that returned nothing.
The agents give you leverage because you understand the work well enough to verify it.
Make every city easier to launch than the last
You will decide which repeated steps should become scripts, agent workflows, product features, or clearer operating practices.
You will test those improvements on real implementations and work closely with engineering to bring the best ones into the product.
You will also bring what you learn from city staff, builders, and homeowners back into product decisions. You are not simply delivering an existing product. You are helping shape what Govstream.ai becomes.
You are our first Customer Success Engineer, not our last. The people who follow will use the playbook, tooling, and standards you create.
The work will be messy. City data will be incomplete. Staff may disagree. An agent will sometimes produce an answer that sounds plausible and is wrong.
When something is unclear, you ask the right questions and work with the city to reach a clear, shared understanding.
You will own the outcome, but you will not work alone. Founders and engineers stay close to every launch, and you will have direct access to the people building the product.
At the end of the work, people across a city will use what you launched. The methods you build will shape every city that follows.
In any given week
You might:
- Direct agents through years of permit records, inspect what they extracted, and trace a wrong answer to a missing GIS response or flawed rule.
- Sit with permit reviewers who disagree about when a drainage plan is required, walk through a real case, and help them reach a clear decision.
- Turn a submittal checklist, zoning table, and portal walkthrough into a defensible specification while naming the cases that remain ambiguous.
- Write a backtest report, walk city staff through the evidence permit by permit, and identify what needs to improve next.
- Guide builders and homeowners through their first live applications, then bring what confused them back to the product team.
- Catch a chart that is technically correct but misleading before it reaches city leadership.
- Turn a manual step from this week's onboarding into a script or agent workflow for the next city.
- Work through the next city's rollout strategy with the founders and the team, and change the playbook based on what the last city taught you.
- Run the weekly city check-in, give a clear account of progress, and raise a risk before the city has to ask about it.
What you'll bring
This is a customer-facing engineering role. The technical bar is real, and so is the customer bar.
We do not expect you to know permitting or House before you join. We do expect technical fluency, careful judgment, and a strong sense of ownership.
You:
- Work effectively with AI coding agents and check their work rather than accepting plausible output at face value.
- Are fluent in SQL and comfortable moving between SQL Server and Postgres.
- Write working Python and are comfortable in a terminal. Git, Bash, and jq are normal tools for you.
- Think clearly about boolean logic, thresholds, types, edge cases, and failure modes. You can pick up a small new language quickly.
- Can own a data integration end to end: extract data from a difficult system, reshape it, load it, and verify that it arrived correctly.
- Debug from evidence. You read logs, traces, payloads, and test failures before forming a conclusion.
- Turn messy documents and conversations into clear specifications. You name ambiguity instead of smoothing it over.
- Can lead a business-process discussion, keep the room productive, and help people reach a decision.
- Stay clear about what is a confirmed city rule and what is still Govstream.ai's interpretation.
- Write clearly for customers and colleagues. Your status reports remain accurate even when the news is uncomfortable.
- Advocate on two fronts: for the customer's outcomes, so permitting actually becomes faster and simpler, and for the people using the product, so city staff and applicants feel informed, capable, and in control.
- Can own a customer relationship over months, keep commitments visible, and make hard calls when needed.
- Are curious, self-directed, dependable, and generous with the people you work with.
- Are someone colleagues and cities genuinely enjoy working with.
Valuable, but not required
- Experience in permitting, planning, local government, construction, architecture, civil engineering, or computational design.
- Implementation or delivery experience in govtech, vertical SaaS, or another regulated field.
- Experience with Accela, Tyler, OpenGov, or similar systems.
- GIS experience, including ArcGIS, QGIS, or spatial SQL.
- JavaScript, for reports and internal tools.
- Elixir or Phoenix. The Govstream.ai platform uses both, and being able to read the code helps.
You might be
- A forward-deployed or solutions engineer who has done real configuration and data work, not just demos.
- An implementation engineer from a regulated field who treats the customer's rules as the specification.
- A data engineer who enjoys working directly with customers.
- A GIS analyst or planner who codes.
- An architect, civil engineer, computational designer, or construction professional who codes and has experienced permitting from the applicant side.
- A permit technician or plan reviewer who moved into technology.
Your background can vary. The same three things should be true: you can build, you tell the truth about the data, and cities trust you in the room.
How we evaluate
Our interview process is practical and based on the work itself.
Expect exercises using real, public city material. You might turn a permit checklist into a specification, triage failing test cases, find the trap in a permit-data export, or work through a configuration change using Claude Code.
We care much more about how you reason, verify, and communicate than whether you have used our exact tools before.
How we work
Govstream.ai is a small, ambitious, remote-first team. Everyone works close to customers, builds, and ships.
We share work early, review one another's thinking, communicate directly, and take ownership when something goes wrong.
We use AI heavily, especially Claude, but people remain responsible for every decision and every review. We value speed, craft, candor, and keeping our commitments.
We avoid process for its own sake. We move quickly without treating quality or trust as optional.
Traction
Govstream.ai is live in Bellevue and Louisville, with active pilots in eight cities.
We are well funded and at an important point in our growth. By Q4 2026 and into Q1 2027, we expect to have 15-20 cities on the platform.
Our ambition is to make permitting work better in thousands of cities, in the United States and around the world.
Govstream.ai is still early enough that you will shape not only your own function, but the company itself.
Details
- Location: Remote-first and US-based. Greater Seattle Area preferred, but not required.
- Compensation: Competitive salary and meaningful early-stage equity.
How to apply
Email careers@govstream.ai with your resume or CV and a one-pager on this prompt:
Tell us about a customer engagement that changed how you work. It might have been one that was going wrong or harder than anyone expected, or simply one where something told you the way you were working wouldn't hold: too slow, too manual, too dependent on you, or unable to scale. We want to know what the signal was, what you did about it, and what you built as a result: the process you designed and the automation you put around it so it ran better the next time.
You can answer in writing or by video, whichever suits you, and a page or a few minutes is plenty. We care most about how you think, so tell the story in your own words, and include links or examples if they help.