Aditi Avasthi

Founder · AI builder · systems thinker

Give me a vast, difficult problem with a human being at its centre.

I built AI for mass-market education years before AI became the obvious thing to build. Now I’m bringing scarce engineering judgment to the businesses and workers who need it most.

What has me now
Aditi Avasthi smiling outdoors
Builder. Writer. Perpetually curious.
01 / Now

AssetBlue

Building intelligence for the physical world.

AssetBlue is a hard technical problem: building a world model for industrial machines that combines deep physics with the ability to reason, without turning either into a black box. The aim is to make scarce engineering judgment available to the operators, technicians and maintenance teams closest to the machines, so they can understand what is happening, what may fail and what to do next.

Visit AssetBlue

Alongside

Ideas I keep returning to.

Soulware

AI infrastructure for human outcomes: memory, context and interfaces that help technology perceive, understand, reason, guide and grow.

AI × human potential

Nizam

An exploration of bespoke intelligence, craftsmanship and the systems that make technology feel deeply personal.

Technology × craft
02 / Embibe

Founder & former CEO

I built AI for one of the world’s hardest mass markets.

In 2012, AI was far from fashionable. I founded Embibe to answer a difficult question: could technology understand every learner closely enough to help them fulfil their potential? We built the intelligence from first principles, then took it across consumers, schools, institutions and public education systems.

  • 190M+people reached across the Embibe ecosystem
  • 35M+learners served through institutional deployments
  • 14,000+institutions served
  • 697curricula mapped around the world
Students using an Embibe interactive lesson together
The point of the technology was always the learner.

The intelligence underneath

Embibe grew into a computational model of learning. Its knowledge graph connected more than 138,000 concepts, 414,000 competencies and nearly 1.9 million explicit relationships.

More than three million assessment items were tagged across 697 curricula and examinations. A single question could carry up to 63 dimensions of metadata, allowing the system to distinguish a knowledge gap from a guess, a timing problem or a behavioural pattern.

That same intelligence layer connected teaching, learning, practice, assessment and counselling across 12 languages. It could absorb a new examination in days and turn each answer into a diagnosis, then a personalised path forward.

Embibe teaching interface with an interactive atomic model
One system for teachers to create, teach, assess and respond.
Embibe immersive virtual chemistry laboratory
Virtual laboratories made the invisible explorable.

The Embibe story

Thirteen years of building toward a no-regrets future for every child.

Moving on from Embibe

Tagging more than three million assessment items across 697 school curricula worldwide against Bloom’s taxonomy revealed something uncomfortable: education overwhelmingly rewards recall and basic application, the capabilities AI can replicate most easily.

We create new frameworks. We empathise. We question. We imagine, contradict, care and make meaning.

Education has to develop the deeply human capacities through which people can outperform AI. That realisation made me far more interested in what technology is for.

It is the question I carried with me when I moved on from Embibe. It continues to shape what I build, what I write and what I am trying to understand now.

Recognition

A decade of building in public.

Recognition for technology, education, entrepreneurship and the work of expanding access to intelligence.

Selected media

Conversations, profiles and ideas.

Forbes IndiaEmbibe: A teacher’s pet toolProfile · 2015Antler VC CastReimagining EdTechPodcast · 2020World Economic ForumAditi Avasthi SurriFounder profileWorld Economic ForumCreating impactful changeLeadership · 2023Economic TimesLeadership is more about EQ than IQConversation · 2019YourStoryThere is so much talent and hard work everywhereInterview · 2014
Four Letter Words

03 / Writing

The unruly human forces underneath everything we make.

Four Letter Words explores urge, fear, love, work, rage, hope and the things we learn when life refuses to stay professionally assembled, through my own life and the lives of people I find impossible to ignore.

Read Four Letter Words
Featured essay · Aug 03, 2026
URGE

The body knows before the business plan does.

Entrepreneurship begins with ownership: the moment you stop looking for someone else and accept that if something stinks, it’s yours to fix.

Read the essay
04 / About

I care about what intelligence makes possible, and who gets to use it.

My work has moved from banking to education to the physical world. The thread is access: who gets expertise, who gets to act, and what changes when intelligence stops being scarce.

Before founding Embibe, I led digital initiatives for Barclays across Africa and the United States and helped establish TCS’s first office in Scotland. I studied electronics at Thapar University and business at Chicago Booth, with executive leadership training at Harvard Kennedy School.

I build because large, difficult problems make me come alive. I write because the emotional reality underneath that work is often where the most useful truth lives.

LinkedIn Substack Google Scholar