top of page

Merly Raises Seed Round for AI-Driven Code Quality Systems at Repository Scale


July 18, 2024


Palo Alto, CA – Merly.ai, the leader in repository scale code quality and understanding, today announced its $6.8 million Seed Funding, co-led by Greycroft and IBM Ventures, with participation from JAZZ Venture Partners. Merly will leverage this capital to further develop the world’s best code reasoning systems.


The proliferation of AI-generated code presents an increased risk of software vulnerabilities and technical debt. The way to address these concerns begins with a deep understanding of the codebase. Merly’s AI platform, Mentor, is the first-ever system that can reason about software quality across a code repository’s entire lifetime; a feat previously believed to be computationally intractable. Trained on one trillion lines of high-quality open-source code, Mentor is Merly’s first step in its larger vision to redefine the way software is developed using AI.


One of Mentor’s technical differentiators is its ability to perform deterministic code reasoning with a novel AI architecture invented by Merly's founder, CEO, and Chief Scientist, Dr. Justin Gottschlich. Justin spent the last 25 years in industry and academia, where he engineered and researched large-scale software and AI systems. Before founding Merly, Dr. Gottschlich spent a decade at Intel Labs where he helped establish the field of Machine Programming (MP). Justin continues to pioneer public domain MP technology at Stanford University, where he teaches and conducts research.


“Mentor provides the much-needed code quality system of the future, which we believe will improve the efficacy, reliability, and adoption of AI coding assistants,” says Dr. Justin Gottschlich. “Engineers spend 70% of their time trying to understand code vs. writing it. LLM-based solutions are great for simple code generation tasks but can often misinterpret or misunderstand larger-scale code semantics. Mentor focuses on code understanding, which we believe is not only more useful from a developer productivity perspective, but also imperative for many code characteristics, such as security, modularity, and performance. With Mentor, engineers and executives, alike, can monitor the quality of closed- or open-source repositories over time—a persistent software engineering requirement made only more necessary by the advent of AI-generated code.”


“We are delighted to see Merly’s code reasoning platform – Mentor – introduced to the market. Mentor’s ability to digest complex code, understand critical dependencies, and leverage proprietary models to self-maintain the code over time is the result of several scientific discoveries. Mentor has already ingested and run inference on billions of lines of large enterprise code—code which critically supports the software and hardware products consumers use every day,” says Brentt Baltimore, Principal at Greycroft.

 

“We are excited to invest in Merly, whose AI we believe will leapfrog code quality tools previously based on simple, hand-crafted rules,” says Roger Premo, General Manager of Corporate Strategy & Ventures at IBM. “Mentor’s ability to scan and detect errors across an entire repository has enormous implications for code modernization and complements our efforts at IBM helping clients maximize the value of their technology investments.”


Mentor is already being used by thousands of early adopters and is performing over a billion lines of code inference per week. Justin and his team have begun establishing collaboration with the world’s best developers, including those at IBM and the Cloud Native Computing Foundation (part of the Linux Foundation). To apply for early access to Mentor, please visit the Merly website: www.merly.ai/early-access.


1 Comment


Congratulations to the Merly team on this great accomplishment. Super excited to see what is to come. Surely this team is going to continue to break ground at Merly speed!

Like
bottom of page