Sonarly Launches 'AI Engineer' to Automate Bug Fixes
A new tool called Sonarly has launched as an "AI Engineer for Production." The platform uses AI agents to autonomously turn production alerts into bug fixes. The service is aimed at backend and AI developers to help reduce the time spent on resolving production issues.
- The company, a member of YC's W26 batch, was founded by Dimittri Choudhury and Alexandre Klobb, who previously co-founded and scaled an edtech app, Meoria, to over 100,000 users in France before dropping out of their computer science program. - Sonarly's technical approach is to solve the "context gap" that limits code-generating AI; while models are effective at writing code (build time), they fail at fixing production issues (run time) because they lack situational awareness from logs, metrics, and traces. - The platform's workflow reflects a multi-agent system pattern, where distinct, specialized AI agents could handle different stages of the software development lifecycle, such as a "triage agent" for alerts, an "investigator agent" for root cause analysis, and a "coder agent" for generating the patch. - From a backend architecture perspective, integrating a tool like Sonarly requires an API-first design with robust observability, exposing production data through platforms like Sentry and Datadog and communication channels like Slack and Discord for the AI agent to consume and act upon. - In an insurtech context, such an AI engineer could maintain the reliability of automated claims and underwriting pipelines; reducing Mean Time to Repair (MTTR) on these core systems is critical, as downtime or bugs can directly impact risk assessment and settlement speed. - The venture capital landscape for insurtech has seen funding stabilize around $1.1 billion per quarter, with investors becoming more selective
