Papaya, a startup developing an optimization platform for AI agents, has joined Y Combinator’s Fall 2026 batch as it works to help companies improve the performance, cost and latency of AI-powered workflows. Y Combinator lists Papaya as an active Fall 2026 company based in San Francisco. The startup has three founders: Macy Mody, who serves as CEO; Niyaz Puzhikkunnath, who serves as chief product officer; and Faiz Vadakkumpadath.
The company’s product analyzes production traces generated by AI agents to identify problems affecting performance. Papaya says it can provide recommendations and turn approved changes into pull requests, allowing engineering teams to address issues without manually reviewing production traces.
According to Y Combinator, Papaya connects to an AI agent through its software development kit and analyzes areas including context, prompts, prompt caching, subagents and tool calls. The company says it runs more than 200 research-backed analyses and ranks recommendations by their potential impact.
READ: Former TikTok employees launch AI app to help users pose for better photos (September 15, 2026)
Papaya’s founders said in a LinkedIn announcement that they had been building the company together recently before publicly announcing its acceptance into YC’s Fall 2026 batch. Mody said the company is focused on helping engineering teams identify why an AI agent may become slower, more expensive or less accurate in production. Instead of requiring engineers to manually examine traces, Papaya is designed to automate that analysis and surface potential improvements.
The startup’s approach centers on production data generated by agents rather than solely on pre-deployment testing. Y Combinator describes Papaya as an AIOps, infrastructure and enterprise software company focused on continuous optimization for AI agents.
Y Combinator says Papaya’s analyses can cover issues such as context bloat, excessive subagent activity and broken tool calls. Its company profile also says recommendations are accompanied by the production runs that generated them and that the startup has seen a typical 10% or greater quality improvement during an initial workflow analysis. Mody previously led go-to-market and operations at SafeBase, where she worked from the company’s early stage through its acquisition, according to Y Combinator. Puzhikkunnath formerly worked as a senior software development engineer at Amazon on machine learning systems, while Vadakkumpadath spent about a decade at Amazon as a senior data engineer, according to the YC profile.
READ: Cognition raises $2 billion as AI coding startup hits $48 billion valuation (September 9, 2026)
The company enters YC as businesses increasingly deploy AI agents for tasks that require models to interact with software, tools and internal systems. As these systems operate in production, companies face challenges around reliability, response time and the cost of repeated model and tool calls.
Papaya is now part of Y Combinator’s Fall 2026 cohort and is continuing to develop its platform for companies seeking to optimize AI agents after deployment.