We build AI estimating software for commercial HVAC, plumbing, and electrical contractors. Before bidding on a job, a contractor or estimator has to count every duct, fitting, diffuser, and piece of equipment across a set of construction drawings that can run to thousands of pages. We speed this up by ingesting the drawings, detecting the equipment schedules, finding symbols, and producing a priced takeoff that the estimator reviews and corrects in a collaborative canvas editor. We have strong PMF and raised a $14M Series A earlier this year.
It may not sound sexy, but it actually is!! There are a ton of hard and interesting engineering challenges. Just to highlight a few:
- very dense document drawings with a realtime collaborative canvas editor (think 50k DOM elements and up)
- segmentation models (and many more) whose output has to become geometry a user can interact with
- gpu inference across hundreds of concurrent jobs
- dynamic analytics and chart generation from agentic workflows custom built for estimators
Open roles
- Software Engineer, Product ($180k–$240k): * Software Engineer, Applied AI ($150k–$250k): * Software Engineer, ML Infrastructure ($200k–$260k): * Software Engineer, Data Platform ($170k–$220k): Stack: TypeScript/React/Next.js, Python, PyTorch, Postgres, Temporal, AWS, Kubernetes, Modal.
We're also hiring for PMs, implementation, sales, and marketing roles! I'm the Head of Engineering. Apply through Ashby and mention HackerNews or HN in the Ashby form. Or email me at john@withrebar.ai. We will reply to everyone, but it might take us some time! Thank you in advance for your patience
Additionally, there's a fun puzzle on our marketing site that will fast track your application if you want to solve!