Skip to main content
Peeyush Ranjan, ex-Google executive, launches AI EdTech Fermi.ai in India and USA, speaking at a podium. [kdesign.co](https:/

Editorial illustration for Ex-Google Exec Peeyush Ranjan Launches AI EdTech Fermi.ai in India, USA

Ex-Google Exec Peeyush Ranjan Launches AI EdTech...

Ex-Google Exec Peeyush Ranjan Launches AI EdTech Fermi.ai in India, USA

3 min read

Peeyush Ranjan, who spent years steering technology at Google and most recently as its chief technical officer, is betting on a different kind of tutoring model. Earlier this month he unveiled Fermi.ai, an AI‑driven education startup that will operate out of both India and the United States. The service promises to move beyond the typical “answer‑on‑demand” approach that dominates many online learning tools.

Instead of handing students a final result, the system watches each move they make on a problem, flags where their logic stalls, and then walks them through the missing pieces. Ranjan argues that this shift matters because the speed at which answers are now available often outpaces the depth of comprehension students actually develop. The platform uses AI to analyse how a student arrives at an answer, identifying gaps in reasoning and offering step‑by‑step guidance instead of direct solutions.

“Students today are getting answers faster than ever, but their understanding is getting weaker,” Ranjan said, who has also served as CTO.

The platform uses AI to analyse how a student arrives at an answer, identifying gaps in reasoning and offering step-by-step guidance instead of direct solutions. "Students today are getting answers faster than ever, but their understanding is getting weaker," Ranjan said, who has also served as CTO of Flipkart and held leadership roles at Airbnb. "We built Fermi.ai to support thinking, not replace it, and to give educators visibility into struggles that usually stay hidden." The startup has emerged from Meraki Labs, where Ranjan partners with entrepreneur Mukesh Bansal, founder of Myntra.

Bansal said Fermi.ai focuses on mapping a student's thought process rather than solving problems for them. "It's about showing students how they think, and helping teachers guide them back to mastery," he noted. Bansal also launched Nurix, which is another emerging AI startup focused on building AI-native consumer and internet products.

Fermi.ai's platform is built around four core components: an adaptive real-time tutor that provides pedagogically guided hints, a stylus-first smart canvas designed for handwritten equations and diagrams, a concept-linked question bank aligned with exams such as AP, IB and JEE, and an analytics layer that pinpoints where student reasoning breaks down. Before its public launch, the company conducted a three-month pilot with 79 students, covering over 15,000 concept tests. According to the startup, students who initially struggled showed an average improvement of 4.68 points by their final attempts, while heavy users demonstrated higher mastery gains and reduced dependence on hints.

The cloud-based platform is currently available for free at fermi.ai, with a dedicated pilot programme for educators launching in 2026.

Fermi.ai enters the market with a clear premise: AI can scaffold the “productive struggle” that Ranjan believes is missing from today’s fast‑answer culture. The Singapore‑based firm has already set up subsidiaries in India and the United States, and its initial catalog covers mathematics, physics and chemistry for high‑school students. By analysing each step a learner takes, the platform claims to spot reasoning gaps and provide guided hints rather than outright answers.

Ranjan’s background as a former Google GM and VP lends credibility, and his comment that “students are getting answers faster … but their understanding is getting weaker” frames the venture’s mission. Yet, the actual impact on comprehension remains to be measured; adoption rates, teacher integration and long‑term retention data have not yet been disclosed. Moreover, the scalability of a model that relies on detailed step‑by‑step AI feedback across two very different education systems is still uncertain.

The company’s next moves, including any partnerships or performance studies, will determine whether the approach can move beyond promise to demonstrable results.

Further Reading