One company,
the whole stack

OBRI Private Limited is an education technology company incorporated in India in July 2026.
We build everything a school needs to run an artificial intelligence programme.

Designed as one system

We design and build the processing hardware, the sensor ecosystem, the experiment curriculum and the learning platform that ties them together. Because all four are built in house, they are designed as one system rather than assembled from parts that were never meant to meet.

We are a product company, not a build shop. Our vision: every student should leave school having built something intelligent with their own hands. Not a simulation, not a video, not an app that calls somebody else's model — something that senses the world, decides, and acts.

Smart AI Box 2.0 processing unit

Four capabilities, one delivery

Hardware design

Smart AI Box 2.0 and a ten-box modular sensor ecosystem — enclosure, board, firmware and peripheral integration, designed in house. Edge AI inference for vision, speech and sensor models, with no cloud dependency.

Curriculum development

Guided, outcome-mapped experiments across Grades 5 to 12, written against syllabus requirements rather than bolted on afterwards. Over 200 experiments across the box and sensor ecosystem.

Platform engineering

Luugo, our AI-powered LMS: assignment, assessment, mastery tracking and adaptive sequencing for every student. Automated assessment generation and personalised learning paths.

Deployment and training

Teacher onboarding and lab setup, so a school can run the full programme without hiring a specialist.

Why inference has to be local

  • Cost. Per-seat cloud AI pricing does not survive a whole-school rollout.
  • Bandwidth. The schools that most need this have the least reliable connectivity.
  • Privacy. K-12 student data should not leave the building to reach a model.
  • Pedagogy. A hidden API endpoint teaches nothing. A model you can open does.

Every model on the Smart AI Box runs on the device itself. Our research track on the NVIDIA Jetson Orin Nano extends this to small language models, CUDA and TensorRT accelerated, entirely offline.

The people building it

Nick Varghese

Nick Varghese

Chief Executive Officer

Sets product direction and commercial strategy; background spanning AI systems, data and curriculum design.

Dr. Saju Varghese

Dr. Saju Varghese

Technical Advisor

Advises on engineering direction, architecture decisions and technology partnerships.

Krishnan Unni

Krishnan Unni

Head of Research & Development

Leads mechatronics and embedded software; previously delivered fielded humanoid service robotics.

Nitin Khan

Nitin Khan

R&D Engineer

Firmware, sensor integration and the Jetson Orin Nano research build.

Turn a classroom into a working AI lab

We are looking for founding partner schools, distribution partners and research collaborators.
Contact us