AI research startup Gibran raises $2.6 million seed funding from Together Fund

AI research startup Gibran has raised $2.6 million in seed funding from Together Fund, co-founded by Girish Mathrubootham (freshworks) and Manav Garg (Eka Software).  

The funds will be used to build basic AI models that work well with small datasets. The first focus will be on drug discovery, where large datasets are hard to get because of limited patient numbers.

Read also- IPO-bound AI company Fractal secures $170 million at $2.44 bn valuation

“Our central thesis is that AI should not replace humans, but rather evolve with them,” said Balakrishnan in a statement. “That means building systems that learn not just from data, but from people—continuously and contextually.”

Manav Garg from Together Fund said these AI systems could transform fields like biomedicine, filmmaking, design, and education. “It’s not just about boosting productivity—it’s about helping people thrive,” he added.

About Gibran

Founded by Govind Balakrishnan and Srikant Chakravarti, Gibran wants to change how machines work alongside humans in creative fields like drug discovery and education.

The team includes Balakrishnan and Chakravarti, who earlier co-founded Curio—an AI-based audio news app that ran for eight years before closing in January 2025. They are joined by Sadedin, an evolutionary biologist using AI to study thinking and Duéñez-Guzmán, a researcher in generative systems.

Advertisment

Gibran is focusing on areas like drug discovery and theoretical science, where there aren’t many labeled datasets. Its AI agents are designed to mix known ideas in new ways—helping humans come up with fresh hypotheses or molecular designs, not just automating tasks.

Instead of using traditional AI models that depend on huge training datasets and fixed rules, Gibran is building “scale-free” models—flexible systems that learn and evolve over time through user interaction.

Together Fund stated that the research will focus on AI that can learn continuously, stay ethically aligned, and act independently. These models could eventually assist in science, creativity, and personalized learning.