Platform Engineer

Underdog.io -Apply to top tech jobs in 60 seconds. A place where companies apply to you; New York (NY); On-site; USD 180,000 - 240,000; Full time

Job summary

Underdog.io is seeking a Platform Engineer to own data ingestion pipelines, AI workflow orchestration, and core platform primitives for government-focused software. This foundational role builds the layer that downstream products rely on.

You will work with the founder to ship reliable data pipelines from mainframe exports to modern storage, while ensuring security and auditability with a path toward FedRAMP/CJIS compliance. US citizenship required; on-site in Flatiron.

Qualifications

• Real systems built from blank page to production with constraints.
• Portfolio and shipped work matter more than pedigree.
• Strong backend fundamentals in Python/Node with API design.
• Frontend capability to reason about an interface.
• Daily use of AI coding tools as part of work.
• Good instincts for abstraction and platform primitives.
• US citizenship required for government work.
• On-site five days a week in Flatiron.

Responsibilities

• Own data ingestion, pipelines, and normalization layer for mainframe exports and APIs.
• Build AI workflow infrastructure for long-running, auditable workflows.
• Develop and maintain core platform primitives including a workflow engine and RBAC.
• Enforce security and compliance by default with data isolation and access controls.
• Collaborate directly with the founder on architectural decisions.

Skills

Python

Node.js

API design

Data modeling

AI coding tools

Frontend basics

Tools

Claude Code

Codex

Gemini CLI

Factory

Job description

A venture-backed, early-stage software company is building AI-native operating software for state and local government. Founded by a repeat founder following a successful exit, the company is already generating meaningful and accelerating revenue.

State and local government is a $100B+ software market that modernization largely skipped. Legacy vendors designed their systems for Fortune 500 enterprises. Defense-oriented platforms were built for agencies with effectively unlimited budgets. Meanwhile, thousands of counties, municipalities, school districts, and utilities deliver essential public services on infrastructure that predates the commercial internet. This team is building the platform those agencies should have had all along.

The Opportunity

This is a foundational engineering role, not a feature assignment. The Platform Engineer owns the layer everything else is built on: data ingestion pipelines, the AI workflow orchestration engine, and the configurable primitives that hundreds of downstream products depend on.

Those downstream products are assembled by Forward-Deployed Product Managers who embed directly with city and county teams — a pretrial release system for a county court, a helpdesk for a school district, an operations dashboard for a municipal water utility, a lightweight ERP for a mid-sized city. The FDPMs configure and ship those solutions. The Platform Engineer makes that possible.

The problems are genuinely unsolved. How does a modern platform pull clean, reliable data out of a mainframe that has been running since the Reagan administration? How do multi-step AI workflows stay auditable enough for a public agency to defend them in a records request? How do primitives stay expressive enough to model any government process while remaining simple enough for a non-engineer to build on?

Whoever answers those questions well will shape how government software works for the next decade.

What the Role Involves

• Ingestion and data modeling. Designing and owning the connectors, pipelines, and normalization layer that turn mainframe exports, Oracle databases, flat files, and proprietary vendor APIs into a consistent, queryable foundation.
• AI workflow infrastructure. Building the machinery behind long-running, multi-step agentic workflows — task queuing, state management, retry semantics, human-in-the-loop approval gates, and the audit trails public-sector compliance demands.
• Core platform primitives. Building and maintaining the workflow engine, role-based permission model, form and action framework, and event system that let a non-engineer stand up a permitting portal or case management tool without rebuilding the plumbing each time.
• Security and compliance by default. Treating tenant data isolation, granular access control, and end-to-end auditability as first-class platform features, with a deliberate path toward FedRAMP, CJIS, and HIPAA certification as the company scales.
• Architecture with the founder. Partnering directly with the CEO on decisions that compound across every deployment the company ever ships.

What the Team Is Looking For

• A track record of building. Real systems, taken from blank page to production, ideally under real constraints. Portfolio and shipped work matter more than pedigree.
• Strong backend fundamentals in Python and/or Node, with sound judgment around data modeling, API design, and building systems that remain reliable and extensible as requirements shift.
• Enough frontend capability to build and reason about an interface, even though this role will not own the UI.
• Daily, hands-on use of AI coding tools — Claude Code, Codex, Gemini CLI, Factory, or similar — as a genuine part of how the candidate works, not an occasional experiment.
• Good instincts for abstraction: knowing when a pattern deserves to become a platform primitive and when it should just ship.
• US citizenship, required for government work.
• Willingness to work on-site in Flatiron five days per week.

Nice to Have

Experience building developer platforms or internal tooling; devops and infrastructure depth; agentic AI orchestration; or prior work in a regulated environment such as healthcare, financial services, or legal.

This role sits on a small, in-person engineering team in Flatiron. Compensation includes a base salary plus meaningful early-stage equity.