PermitFlow
Applied AI Engineer
About this role
PermitFlow is seeking an Applied AI Engineer to develop and manage AI agents for their pre-construction platform, enhancing efficiency in construction workflows. The role encompasses the whole lifecycle of AI agent deployment from development to evaluation, contributing to a transformative shift in the construction industry.
What you'll do
- Develop and deploy AI agents for automating permitting workflows
- Orchestrate models to create reliable AI systems
- Build internal tools for agent monitoring
- Design backend and frontend services and APIs
- Contribute to system architecture with the engineering team
- Optimize pipelines for speed and scalability
What they're looking for
- 3+ years in software or ML engineering
- Experience with production AI/ML systems
- Proficiency in building multi-step ML systems
- Strong backend and API development
- Attention to quality and detail
- Comfort in fast-paced environments
- Some frontend development experience
- Startup grit
Benefits
- Competitive salary and equity
- Full medical, dental, and vision coverage
- 401(k) savings plan
- Unlimited PTO and paid family leave
- Home office stipend
- Daily in-office meals
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
PermitFlow
PermitFlow builds AI-driven software solutions that streamline permitting and pre-construction workflows in the construction industry. The company is hiring engineers across design systems, full-stack development, applied AI, customer deployment, and analytics to enhance its platform's capabilities and user experiences.
- Website
- permitflow.com
Likely interview questions
- Walk us through a production AI or ML system you shipped end-to-end. What was your role, and what challenges did you face in moving it from development to production?
- Describe your experience building multi-step AI workflows or agentic systems. How did you handle orchestration, error handling, and reliability across multiple model calls?