The role
- COMP
- $165K - $270K
- EQUITY
- percentage to be decided as part of offer
- LOCATION
- New York
- WORKPLACE
- On-site
- EXPERIENCE
- 3 - 6 years
- VISA
- None
- STACK
- REST API, PostgreSQL, Vector Database, AWS, GCP, Docker, TypeScript, CI/CD
- INDUSTRY
- AI, Insurance
/roles — ROLE_367
Founding Software Engineer
Accelerator-backed insurtech startup whose AI agents handle calls and paperwork in insurance claims
The company
An early-stage startup using multimodal AI agents to automate insurance claims work, taking routine phone calls and document processing off adjusters' plates at large carriers.
- STAGE
- Series A-stage
- FUNDING
- $5M raised
- TEAM
- founding team of <10
- FOUNDED
- 2025
JD — the work
About the role
You would fully own the central platform behind the company's AI agents, which already run in production with several of the largest insurance carriers while revenue grows very quickly. The team is under ten people, backed by a top accelerator, and releases to production weekly. Expect little hand-holding: you architect systems, coordinate agents, and own the reliability of software that customers depend on.
What you'll do
- Own full platform components: model reliability, evals, document parsing, and the infrastructure agents run on
- Add resilience features like failover, circuit breakers, retry logic, canary releases, and observability
- Create an evaluation platform with evals on every run so regressions are caught before customers see them
- Work closely with forward deployed engineers to spec and ship fixes for real customer pain points
- Improve developer tooling, CI/CD, and coding-agent integrations to speed up the team
What they're looking for
- 4+ years of engineering, primarily in TypeScript and SQL in production, clearly shown on your resume or GitHub
- Hands-on experience running production systems on AWS
- Startup experience (Seed to Series B) where you owned a central piece of the platform, not only features
- Background in fintech, insurtech, healthtech, or another regulated, security-sensitive field where auditability was mandatory
- A public artifact, such as open-source work, a personal site, or a publication, showing engineering curiosity outside work