Software Engineer, AI-Native

About CityScout

CityScout is a real estate development intelligence platform built from government public records. We turn zoning cases, building permits, planning documents, parcels, ownership records, and related public data into a living picture of who is preparing to build what, where it is happening, and which companies and people are involved.

Our customers use CityScout to identify opportunities earlier and act with better information. Behind the product is a large, long-running data system spanning web scraping, document processing, geospatial analysis, entity resolution, AI classification, APIs, and customer-facing applications. Our team is remote-first and relies on clear ownership, documented decisions, and intentional collaboration.

The role

We are looking for an early-career engineer who is genuinely AI-native. Coding agents and LLM tools should already be a normal part of how you explore, build, test, and debug software.

One to two years of professional experience is a common profile for this role. The stronger signal is what you can do. We care about whether you can enter an unfamiliar codebase, understand a bounded area, take ownership of a real problem, and ship production-quality work quickly.

During your first week, we expect you to deliver a meaningful production contribution with appropriate support and review. That may be an application component, a scraper, an ingestion workflow, a data-quality improvement, or an internal tool.

What you will do

• Build and ship customer-facing application features
• Develop and maintain web scrapers and ingestion workflows
• Work with geospatial, parcel, permitting, ownership, and project data
• Improve LLM-driven classification and enrichment systems
• Investigate data-quality and production failures across multiple services
• Write tests, documentation, and monitoring appropriate to the risk of the work
• Use coding agents to accelerate implementation, analysis, testing, and review
• Explain your changes, assumptions, tradeoffs, and verification clearly
• Own work through deployment and production follow-through
• Learn how your area connects to CityScout's broader data and product system

What strong performance looks like

• You become productive in an unfamiliar repository quickly.
• You ship a meaningful production contribution during your first week.
• Your work solves the stated problem and handles important edge cases.
• You understand and can explain everything you submit, including agent-generated code.
• You recognize brittle or accidental solutions and improve them before release.
• You communicate blockers and consequential ambiguity early.
• You become increasingly independent while maintaining sound judgment.

What we are looking for

• Demonstrated ability to build and ship production software
• Fluency with coding agents and modern LLM tooling
• Ability to read, understand, and modify unfamiliar codebases
• Strong debugging and technical reasoning skills
• Care when working with production systems and data
• Clear written and verbal communication
• High ownership, good judgment, and comfort operating with incomplete information
• A habit of verifying outputs instead of accepting plausible-looking results

Helpful experience

These are useful advantages, not prerequisites:

• TypeScript, Next.js, or Node.js
• AWS Lambda or other serverless infrastructure
• PostgreSQL, PostGIS, or geospatial systems
• Drizzle or similar database tooling
• Playwright or browser automation
• Web scraping and resilient ingestion pipelines
• LLM classification, extraction, or enrichment workflows
• Real estate, GIS, government records, or municipal data
• MCP servers, agent tool integration, or permissioned AI workflows

Our stack

• TypeScript and Next.js
• Node.js and AWS Lambda
• PostgreSQL and PostGIS
• Drizzle
• Playwright-based scrapers
• AI classification and enrichment pipelines built on current Claude models
• AWS infrastructure across the platform

How we evaluate candidates

The process includes a realistic, time-boxed engineering exercise. You may use the coding agents and professional tools you normally use. We will review both the result and your understanding of it.

We look for speed, quality, AI fluency, technical judgment, ownership, verification, and communication. We place greater weight on demonstrated production ability than on credentials, trivia, or a precise number of years worked.