SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Rebar is building an AI operating system for commercial HVAC, Electrical, and Plumbing. The company's quoting platform has processed tens of thousands of projects across North America, serving many top firms in the industry. With a recent $14M Series A, Rebar is entering a growth phase focused on AI-native products that will transform how the construction industry operates.
You will join the AI and ML team as a backend engineer, working directly with in-house models and frontier VLMs to build world-class user interactions. Your work bridges AI/ML outputs and product surfaces—for example, taking segmentation masks from detection models and determining their optimal representation on the canvas (polygon, polyline, connection ports), designing the data contracts that carry them, and enabling users to interact with, correct, and bill against them.
This is not purely about model improvement; it's about ensuring the ideal shape and use case for what users see, trust, correct, and bill against. You will own the accuracy-versus-UX tradeoff honestly and close the correction loop.
Key responsibilities include:
- Design the perception-to-product data contract
- Ship AI geometry into the canvas
- Own accuracy and UX tradeoffs
- Close the correction loop
- Serve as the translation layer between AI/ML and product engineering teams
You should be comfortable working across the stack—reading PyTorch postprocessing code in the morning and understanding canvas rendering flows in the afternoon. You'll dive into model architectures, spot inefficiencies, and translate outputs downstream.
The role is salaried and onsite in New York City's Flatiron district. Rebar values working together in person for creative problem-solving and strong team connections.
QUALIFICATIONS:
- 6+ years of industry software engineering experience (applicants with fewer years are encouraged to apply)
- Shipped computer vision or ML output into a product surface that users interact with daily
- Comfortable in Python and TypeScript
- Experience designing and versioning data contracts between services on live customer data
NICE TO HAVE:
- Canvas and rendering experience
- Familiarity with vector graphic object models, standards, and protocols
- Experience building human-in-the-loop labeling or active learning loops
- Background in construction, CAD, GIS, or domains where drawings are the source of truth
