Institutes that run well without heroics.

I’m Gunvant Jain, co-founder of Curious Minds - IIT & Medical in Mumbai. I build the people, academic and operating systems that help good teachers become great, and let a multi-centre institute run without depending on any one person, including me.

IIT Madras. Teach For India Fellow. Trained 500+ teachers at InOpen Technologies.

Have a question? Ask the site. It answers only from what is written here.

Now
Co-founder, Curious Minds - IIT & Medical, Mumbai
Studied
B.Tech, IIT Madras, 2010
Began in
Teach For India Fellowship, Pune
Also
InOpen Technologies (SINE, IIT Bombay), Shikshalaya, ICSEPRELIMPAPER.COM
Teaches
Physics, a couple of half-days a week
Before and now at Curious Minds - IIT & Medical
MeasureBeforeNow
Batches per teacher11 batches on 15 teachers, then 18 on 100.71.8
How often the timetable changesEvery weekMonthly or less
Students absent on a given day10–15%2–5%
Seats filled at course end50–75%Over 90%
Time to act on a weak teacher3–4 monthsAbout 1 month

800+ students, 7+ centres, a team of 40+. Figures are from internal records, as I recorded them. Each measure opens the case behind it.

Who I am, in a nutshell

I read people, build systems that make the right behaviour the easy one, and have chosen education as my arena.

At my best

  • A natural mentor. I am at my best when someone I have guided stands on their own. I have trained teachers knowing some would outgrow me. The growth was the point.
  • Self-reflective. When something goes wrong I pause, listen to the people involved, then fix the system, not the person.
  • An adaptive strategist. I change the plan, not the purpose. My early ventures taught me which plans to drop quickly.
  • A long-term builder. I would rather build slowly something that lasts. I have stayed with one institute since 2016, through its hardest years, and I am preparing my own successor.
  • People-smart. I read the room before I try to change it. Every change I have led began with one-to-one conversations.

Where I do my best work

  • Where people matter as much as numbers.
  • Where systems get built, not only run.
  • Where the bond between teacher and student is taken seriously.
  • Where change comes through trust and nudges, not orders, and long-term results count more than quick wins.

What drives me

Real personalisation lives in the bond between teacher and student. That bond only scales when it sits on a well-engineered system. Tools change, from chart paper to AI. The bond does not. Technology’s role is to empower the teacher, never to replace them.

My method is simple. Reflect on the problem alone first. Ask the people closest to it. Decide and deploy. Then explain the reasoning, not only the plan. I nudge rather than mandate.

What I want my work to leave behind

  • Teachers who grew and now lead others.
  • Institutes that run well without me.

I still teach Physics for a couple of half-days a week. It keeps me close to the classroom, and it’s how I stay sharp on what students and teachers actually experience.

My motto

Guide well, build well, then step back.

My vision for my life

To live unhurried and curious, close to my family, and to leave behind teachers who lead and institutes that stand on their own.

Selected work

Six problems I took on at Curious Minds - IIT & Medical: what I did, and what changed. Each case says how its figure was measured.

More batches per teacher, with the teachers on board

A constraint-based timetable, and one-to-one conversations before any announcement, took the institute from 11 batches on 15 teachers to 18 batches on 10.

Batches per teacherfrom 0.7 to 1.8

Internal timetable and roster records, before and after. Unaudited.

Make consistency the goal, and marks follow

A student culture built on turning up on time cut daily absence from 10–15% to 2–5% in four to five months.

Students absent on a given dayfrom 10–15% to 2–5%

Internal attendance and test records over 4–5 months. An association, not a controlled experiment.

Seven ways in, not one

A planned calendar of entry points lifted batch fill from 50–75% to over 90%.

Seats filled at course endfrom 50–75% to Over 90%

Internal admission and batch records. Unaudited.

Hear about a weak teacher before the parents do

A monthly student-feedback loop cut the time to act on a teacher problem from three or four months to about one.

Time to act on a weak teacherfrom 3–4 months to About 1 month

Internal monthly feedback records. A small sample. Unaudited.

Head office holds the money, centres run the classroom

A view-only sheet for every centre, with finance fields controlled by head office, gave centres the data they needed without letting them edit the money.

Misreporting incidents detectedNone in 2+ years

Internal records over 2+ years. No external audit.

Also built

  • A career assessment, years before AI career tools. In 2018 we built two tests, one for inborn problem-solving skills and one for acquired science skills, and a portal that combined a child’s scores with the parents’ own observations to suggest a top five of careers. We had researched 350+ careers and grouped them into about 40 clusters, mapped to 16 skills. Families came in person for a counsellor-led session, which brought whole families to the table. 750+ reports, some offered to charitable schools. We never marketed it. Our own limits held it back, and that taught me something too.
  • Growth Trajectory Predictor. A model that maps a student’s current marks to an expected future result and sets a target for the next test. It was released in 2018 as a website tool, where students entered a Class 8 percentage and saw a projected Class 10 result, and the response was strong. I later used it to set targets for Class 11 and 12 JEE and NEET students after each test, and students became more sincere about tests. I did not capture structured results, so I can’t give you a number. The next version needs a separate curve for each subject, because Maths and Chemistry grow differently.
  • An operating stack with no licence fees. One admission form now feeds both the fee system and the coaching app, so the three systems no longer disagree. Weekly finance oversight fell from 12 hours to 3, and monthly reconciliation errors from about 15 to none. Salary calculation went from two or three days to a few minutes. Test papers went from 2 to 8 a week, with typing errors down from 5–6 per test to 1–2. A looped counselling video in reception raised parents counselled per counsellor from 5–6 to 8–10 a day. Test absence fell from over 15% to under 5%.
  • Dashboards people actually use. A per-student view with cohort percentiles, branch results with notice-board PDFs, and a syllabus check with a Friday routine. I deleted a tracker I had over-built and replaced it with a simple one. Simple won.

Every figure on this site comes from internal Curious Minds - IIT & Medical records, as I recorded them. None has been audited. Each case says how its figure was measured and where it stops being true.

The road so far

In the order I worked on them: Teach For India, InOpen Technologies, Shikshalaya, ICSEPRELIMPAPER.COM, Curious Minds - IIT & Medical, and now LUMEN Academy. In the margin, what was happening in education at the same time.

2010 to 2013

Teach For India, Pune. Two years teaching Grades 3 and 4 in a municipal school, deploying HBCSE’s Small Science curriculum in my own classroom, work the Homi Bhabha Centre recognised. InOpen Technologies (SINE, IIT Bombay). Began in academic sales, moved into teacher training, trained 500+ teachers across India, helped define the post-sale handholding kit for schools, then built a new vertical from scratch around Small Science.

What it taught me. Selling a good product is the easy half. A teacher has to feel confident using it, or it stays in the cupboard.

The Right to Education Act came into force on 1 April 2010, making elementary schooling a legal entitlement. By the end of 2012 the flipped classroom was being named one of the year’s leading ed-tech trends.

2013 to 2016

Shikshalaya (2013). A flipped-classroom coaching centre for school-board students. By 2018 it had taught 350+ students from 12 schools across three boards. ICSEPRELIMPAPER.COM (2015). A digital study-material platform for ICSE Class 10. Curious Minds - IIT & Medical (2016). Co-founded a JEE, NEET and CET institute for Classes 11 and 12.

What it taught me. Ticket size decides reach. A small-ticket product needs very many buyers, so I moved to a higher-touch model.

On 28 April 2016 the Supreme Court ruled that NEET would be the common medical entrance test from 2016, replacing the many separate exams. It reshaped how families chose where to prepare.

2018

Built an aptitude and skills assessment that matched students to careers, and a Growth Trajectory Predictor that set each student’s next-test target. Both started as small experiments and grew into counsellor-led family sessions and a data-led classroom routine.

What it taught me. The best education products do not give answers. They start better conversations.

Eight years later, Maharashtra’s draft coaching law asks every centre to run aptitude assessments and give families realistic feedback.

2020 to 2022

COVID closed classrooms in 2020. Curious Minds - IIT & Medical restarted in 2022 as a single centre.

What it taught me. Systems that depend on one person’s attention are the first to fail under pressure.

Online learning boomed, and Indian edtech funding peaked at about $4.3 billion in 2021.

2023 to 2025

Grew from one centre to seven or more, and built the systems described in the cases below. In late 2025, after a co-founder transition, I took sole charge of academics, operations and finance.

What it taught me. Growth breaks first where everything depends on a single person.

Edtech funding fell to $265 million by 2023. On 18 January 2024 the Ministry of Education issued coaching-centre guidelines: no enrolment under 16, no guarantees of rank or marks, fair fees with receipts and pro-rata refunds, and trained counsellors. PhysicsWallah listed in November 2025.

2026

Rebuilt the timetable, brought in a successor leader so the institute can run without me, and started LUMEN Academy: a small amount of teaching, and a prototype that tests how AI can ask a student the next question.

What it taught me. Guide well, build well, then step back.

PhysicsWallah’s offline centres grew from 198 to 353 in a year, to the end of FY26. Maharashtra published a draft Private Coaching Centres Act on 20 August 2026: separate registration for every branch, fee receipts and refunds, a counselling system, and a five-hour daily cap on classes. AI and computational thinking enter schools from Class 3 in the 2026–27 session.

Where the sector is now

  • Money and attention have moved from apps back to classrooms. Running many centres well is now the hard part.
  • Rules are tightening. Registration, fee transparency and counselling have to be built into how a centre runs, not added afterwards.
  • AI is arriving in classrooms as policy. Teachers need leaders who can bring it in with them, not over them.

How I work

Projects

LUMEN Academy AI

A prototype physics tutor that asks the next question instead of handing over the answer.

Prototype. It exists to show how I think AI should sit in a classroom, and it is not a product for sale.

Ask the site

Ask about my work, how I think, or what a number means. It searches only the text of this site, and it links the pages it matched.

It matches words, so check the linked pages.

Write to me

Message me on LinkedIn about an institute, a role, or a question about anything here. I read every message and reply myself.

Message on LinkedIn

You can also find me on LinkedIn.