Ex-Ramp Engineers Secure $20M Series A for Melius Creative AI Platform

Overview

The artificial intelligence landscape is littered with the ghosts of abandoned codebases, but few pivots are as dramatic—or as lucrative—as the one executed by the founders of Melius. Originally established by former Ramp software engineers to optimize corporate ad spend, the New York-based startup ultimately realized its initial thesis lacked structural longevity. Rather than iterating on a stagnant product, the team took the radical step of incinerating their entire repository and starting from scratch.

That bold gamble has now paid off in the form of a fresh $25 million capital infusion, highlighted by a $20 million Series A round spearheaded by CRV, alongside a prior $5寧 million seed investment led by General Catalyst. Despite stepping out of stealth mode only recently, the restructured company claims to have rapidly crossed the million-dollar milestone in annualized revenue.

Shifting from Optimization to Agentic Generation

The initial iteration of the platform focused on traditional performance marketing—helping brands manage, monitor, and tweak their ongoing ad campaigns. However, after more than half a year of building and deploying, the founders recognized a fundamental market truth: managing workflows was vastly less compelling than automating the actual creation of the assets themselves.

The team redirected their engineering efforts toward the burgeoning world of autonomous systems and multimodal media synthesis. They re-architected their platform into what they describe as an "agents lab for creative work." Instead of parsing analytics dashboards, users interact with specialized multi-agent loops capable of ideating, formatting, and generating ad campaigns, images, and video content directly from natural language prompts.

Navigating a Crowded Multimodal Landscape

Melius enters a fiercely competitive arena populated by both high-flying unicorns and nimble startups, including major players like Higgsfield and specialized tools like Krea and Flora AI. Yet, the leadership remains unfazed by the dense market concentration. They argue that the sheer scale of the global digital advertising market provides ample room for multiple architectural paradigms to thrive.

For co-founder Joowon Kim, whose childhood ambitions included becoming a prominent YouTuber long before entering enterprise software, the mission is deeply personal. By combining deep systems engineering backgrounds with modern generative models, the platform aims to democratize professional video and media production, enabling anyone to execute complex creative direction through plain-text instructions.

Architectural Takeaways

The evolution of Melius highlights a crucial lesson for modern AI startups: domain-specific agentic loops often outperform general-purpose workflow wrappers. By pivoting from passive analytics to active generative creation, the engineering team aligned their product architecture directly with the highest-friction bottlenecks in modern digital marketing.