About Remi

Remi is a Y Combinator-backed roofing technology platform that connects homeowners with vetted contractors across the United States. Founded in 2022 and based in Utah, Remi is modernizing the roofing industry by solving a persistent challenge: homeowners struggle to find reliable contractors, and contractors waste time on inefficient bidding processes. The platform handles everything from instant quoting and contractor matching to project coordination, permitting, and payments.

This complexity demands sophisticated software to orchestrate multiple parties, tight timelines, and detailed project requirements. Remi's engineering team of approximately 25 people faces constant pressure to ship features quickly while maintaining the reliability that homeowners and contractors depend on.

The Challenge

When you need to develop engineers but don't have good data

As Remi's engineering team grew, VP of Engineering Brant Choate faced a problem common to fast-growing teams: how do you help young engineers improve when you can't reliably see what's actually happening?

"I signed up for Weave because I wanted a more reliable way to help our young engineers grow," Brant explains. As Remi's engineering team approached 20 engineers and planned aggressive expansion, coaching and developing talent became critical. But the tools available for measuring performance simply weren't trustworthy enough to base coaching decisions on.

The manual overhead was significant, pulling engineering managers away from actually managing people to instead manage metrics. Story points need constant grooming, JIRA tickets require categorization, and surveys need responses. And after all that work, the data still didn't align with what leadership could see happening on the ground.

The deeper problem was trust; when metrics don't match reality, you can't use them to coach engineers or make important decisions. Brant needed to spot when someone was struggling early enough to intervene, identify top performers to learn from their approaches, and understand where engineers were actually spending their time. But every productivity tool they evaluated produced numbers that felt slightly off or could be easily gamed.

With 25 engineers split across three managers and plans to grow rapidly, Remi needed engineering metrics that were both automated and believable. Without them, coaching became guesswork, performance conversations lacked data, and scaling decisions felt risky.

The Solution

AI native analytics that actually work without the manual overhead

Remi chose Weave for a reason that set it apart from every other tool they evaluated: "The biggest part is it just works. I hook it up to GitHub, and I can get reliable output without forcing a process."

Weave scans code to understand complexity and effort. The platform immediately integrated with Remi's GitHub repositories and began analyzing every pull request, using machine learning models to estimate how long each PR would take an expert engineer to complete, creating a standardized unit of output.

Brant implemented a streamlined workflow where engineers write brief weekly reports in Notion covering their most important goals, then add "additional commentary on top of Weave" to provide context that the metrics might miss. "The combination of the two is giving me quite nice insight into what's going on across a lot of people," he notes.

Weave established reliable baselines for each engineer's standard output patterns, making it clear when something had shifted, whether that was an individual engineer suddenly producing significantly less, or team-wide metrics like bug-to-feature ratios trending in unexpected directions.

The Results

Early warning systems and confident scaling decisions

The most striking early win came from Weave's ability to surface problems before they became crises. "I had an engineer leave. I knew he was going to leave three months ago," Brant recalls. "We even talked about it preemptively. It just becomes obvious when someone is doing 30 a week and then all of a sudden they drop to 10 for three straight weeks."

This early warning system has become Brant's most valuable use case. "I'm probably best using this to see big fluctuations in output. Someone clearly didn't ship for almost a month, what happened?" Sometimes the answer reveals personal struggles or engineers feeling stuck on certain types of work, allowing managers to intervene early and get people back on track.

Weave solved the transparency problem they'd been pushing for. Instead of vague updates based on JIRA ticket counts, Remi can now show actual engineering output trends, AI adoption rates, and resource allocation across projects with metrics the AI engine derives directly from code analysis.

— Brant Choate, Remi