Some problems have no client. We build for them anyway.
The hardest problems rarely arrive with a budget. So we fund our own — putting AI and robotics against the places where systems fail quietly: a school meal that never reaches a child's plate, a feeding program nobody can audit, care that stops when the treatment ends. Built for real deployment, measured in outcomes, not invoices.
Erudex. India’s most teacher-friendly ed-tech platform.
Digital teaching and learning for schools and colleges, grades 1–12 — foundation and competitive programs, virtual labs, augmented reality, assessments and analytics. Built by Indus Valley, used by schools across India.
Blended learning, built around the teacher
A state-of-the-art LMS that lets schools create, manage and deliver lessons — live classes, lesson planning and on-demand recordings, with separate experiences for students, teachers and administration.
- Students — animated video lessons, quizzes and interactive study material, at their own pace
- Teachers — centralised lesson planning, a digital library and tools that cut prep time
- Administration — grades, sections, students and teachers managed from one place
1–12
grades covered
Share study material · live classes · digital content
Grades 1–12
Schools & colleges
5,00,000+
Questions & pre-loaded tests
IIT & medical faculty
Who design the coursework
Since 2014
Erudex Private Limited
One photo. A child’s full meal, counted.
School mid-day meal programs feed millions of children — and where nothing is measured, food quietly goes missing and nobody can prove a child was fed. We are building a vision system that takes a single photo of a plate, with a weighing scale in frame, and returns the dishes served, the grams on the scale, and the nutrition delivered to that child. A measured plate means a child verifiably fed — and a program that cannot be quietly defrauded.
- A fine-tuned vision model recognises the dishes — a closed set of South Indian meal items
- Purpose-built OCR reads the weight straight off the scale display
- Runs offline on edge hardware in the school — children's photos never leave the building
Pulagamrice · moong dal · ghee · cumin · pepper
read straight off the display
Simulation first. Deployment when it’s earned.
Our robotics lab runs industrial-arm simulation and digital-twin experiments on NVIDIA Isaac Sim, Isaac ROS and ROS 2 — proving motion, perception and cell design in software, then closing the sim-to-real gap so it holds on the actual machine.
Industrial simulation
Yaskawa Motoman arm simulation with ROS 2 and MoveIt 2 — motion planning proven in software before any metal moves.
Digital twins
Physically accurate scenes in NVIDIA Isaac Sim, where a cell can fail a thousand times before it is ever built.
Closing the sim-to-real gap
A policy that works in simulation and fails on the floor is the hard problem in robotics. Domain randomisation, synthetic data at scale, and calibration against real measurement — so what passes in simulation survives contact with the machine.
Edge autonomy
Perception and control sized for edge devices — the same discipline as our AI practice, aimed at machines.
Where you’ll find us in person.
Robotics · automation · AI
Products we back and build.
Let's discuss how we can make your business better.
Tell us what you are running and what is in the way. We will tell you honestly whether we are the right people for it.