Skip to content

08. Problem Statement

The Idea

We are building a selective accelerator for high-intent AI and data-domain technical builders in India. Not to teach AI from scratch, and not to run another bootcamp. The goal is simpler and harder: compress the time it takes already capable engineers to become exceptional AI practitioners whose work speaks for itself.

The last generation of upskilling companies optimized for enrollment. That worked briefly. Then trust collapsed. Certificates stopped meaning anything. Placement promises stopped being believed. The market corrected brutally.

We are deliberately choosing the opposite direction. Select fewer people. Go deeper. Optimize for capability first, reputation second, revenue third. If that sounds slow, it is. But institutions are slow to build, however, durable.

The Problem We See

India produces a massive number of engineering graduates every year, yet companies consistently complain they cannot find strong AI and data engineers. This is not a talent shortage. It is a signal shortage. There are three fractures in the system.

First, the talent side. Somewhere in Tier-2 colleges, Tier-3 cities, rented rooms, late-night Discord servers, there are engineers already teaching themselves. They build models, scrape datasets, ship projects, fail interviews, and keep going. These people do not lack intelligence or hunger. They lack structured acceleration, credible signaling, and access to serious peer networks.

Second, the education supply side. Many edtech companies optimized for revenue before outcomes. Aggressive sales, financing tie-ups, placement narratives. Some delivered value, many overpromised. Employers adapted by distrusting credentials entirely.

Third, the employer side. Hiring managers increasingly want demonstrated capability. Real systems built. Real constraints handled. Real collaboration experience. But credible proof-of-work pipelines are rare.

The result is a broken loop where learners distrust edtech, employers distrust credentials, and genuine talent takes longer than necessary to surface. That inefficiency is our opportunity.

What We Are Actually Building

At its core, this is a talent accelerator for AI, ML, and data-domain engineers.

The product is not lectures. Not content libraries. Not certificates. The product is:

  • A highly selective peer cohort
  • Intense project-driven collaboration
  • Practitioner mentorship
  • Credible proof-of-work portfolios
  • A lifelong professional network

Graduation should mean something very simple: If you came through here, employers can trust your capability. That trust is the real product.

The “Proof of Work”

This term gets used loosely, so it is worth clarifying. Proof of work, for us, means work that survives scrutiny outside the classroom:

  • Production-grade AI applications deployed publicly
  • Meaningful open-source contributions
  • Research or technical writing with substance
  • Real industry problem simulations or collaborations
  • Systems built under constraints, not tutorials

Not toy GitHub repos. Not course capstones. Not certificates.

Employers should be able to evaluate someone without reading their resume.

Detailed portfolio standards, evaluation rubrics, and verification mechanisms will be discussed in the FLOATS chapter under Academics and Operations.

Narrowing the Target Market

We are intentionally focusing on the AI and data engineering ecosystem: AI engineers, ML engineers, data engineers, applied data scientists, and MLOps practitioners. Not beginners. Not career switchers initially. Not anyone who wants to code. Primarily, we are looking for

  • Final-year engineering students, often from Tier-2 or Tier-3 institutions, already self-learning and building passionately
  • Early professionals (0 to 3 years experience), with decent profiles

This is a smaller market than generic software upskilling. It is also a better one. Higher intent. Lower churn. Better outcomes. Stronger network effects. We hope, depth builds brand faster than breadth.

Market Size: TAM, SAM, SOM

We prefer realistic sizing to optimistic storytelling.

India graduates roughly 10 lakh engineers annually. Not all pursue software careers. Conservatively assume about 20 percent lean toward software, AI, or data roles. That gives a broad TAM of roughly 2 lakh candidates annually.

Within that, the AI and data domain is still a subset. A reasonable estimate is 40,000 to 50,000 engineers annually seriously exploring AI, ML, or data engineering paths. That is our SAM.

Now apply intent filtering. The people already building, already experimenting, already pushing themselves. That is probably 5 to 7 percent of the above, giving a SOM of roughly 2,500 to 3,500 high-intent engineers annually.

We are not trying to serve everyone. If we eventually serve even a meaningful fraction of that group, the institution becomes significant.

Pricing Philosophy and Unit Economics

The expected price is around ₹1.5 lakh per learner per cohort. This is deliberate. Cheap programs attract curiosity. Premium programs attract commitment. The goal is not exclusion but seriousness. Scholarships remain possible, but selectively and transparently.

Income Share Agreements sound attractive but introduce legal friction, delayed cashflow, and operational complexity. We prefer straightforward pricing initially.

At steady state, a cohort of around 30 learners would generate roughly ₹45 lakh revenue. High-touch mentorship is the primary cost center. Community infrastructure, admissions, and operations follow. Even conservative modeling suggests healthy contribution margins without requiring mass enrollment. Long-term upside comes from employer partnerships, alumni referrals, and brand compounding rather than aggressive scaling.

Detailed financial modeling lives in the FLOATS chapter under Finance.

Program Structure (High Level)

We intentionally avoid rigid curriculum framing. But broadly, the structure looks like this:

  1. Selection and alignment first. Portfolio reviews, technical conversations, motivation checks. We want signal detection, not polish detection.

  2. Project immersion next. Students start building immediately. Lectures exist only when they unblock progress.

  3. Portfolio consolidation follows. Projects become production-grade artifacts, documented and defensible.

  4. Network integration last. Employer introductions, alumni mentoring, ongoing community access.

The exact mechanics are covered in detail in the FLOATS chapter, especially Academics and Operations.

Customer Acquisition and GTM

Early growth will not come from paid marketing. It will come from:

  • Founder-led outreach
  • Technical community presence
  • Alumni referrals
  • Credible technical content

Trust compounds slowly but powerfully. Paid acquisition before trust usually backfires in education. FLOATS Strategy chapter expands this into acquisition economics and sequencing.

Competition: Bootcamps, AI Programs, Self-Learning

We are not pretending alternatives do not exist.

Some Indian bootcamps execute operations extremely well. Structured curriculum, placement pipelines, disciplined sales. Many learners benefit from them. But structurally they remain education providers. We are trying to become a talent institution, closer psychologically to an accelerator or guild than a training company.

AI-specific programs are emerging, but most are still lecture-centric or certificate-oriented. Few have built strong professional communities yet. Self-learning remains the strongest competitor. Online resources are excellent. Many great engineers are self-taught. But self-learning lacks structured peer pressure, credible signaling, and hiring network density.

We do not replace self-learning. We accelerate it.

Why Someone Would Choose Us

Because speed matters. Because network matters. Because signaling matters.

Someone serious about AI can absolutely learn alone. It just often takes longer, involves more trial-and-error, and lacks hiring visibility. We aim to compress:

  • Learning cycles
  • Feedback loops
  • Hiring access
  • Professional identity formation

That compression is the value.

Our Unfair Advantage (Hypothesis)

It is not technology. It is not curriculum. Those are replicable. The emerging advantages are softer but harder to copy.

  1. Founder proximity to the ecosystem. Firsthand exposure to what worked and what failed in the previous edtech wave.
  2. Community-first design. Most programs add community later. We start there.
  3. Selectivity flywheel. Better students attract better students.
  4. Radical transparency. Publishing real outcomes builds trust faster than marketing claims.

Over time, alumni density becomes the moat.

Risks We Acknowledge

Talent discovery is slow. Reputation takes years. Selectivity pressures grow with scale. Founder intensity is initially non-scalable. Community dilution is a constant risk. This is not a blitzscale startup. It is institution building that requires patience and resilience.

Long-Term Vision

If this works, the institution becomes a trusted talent signal. Employers recruit proactively. Alumni mentor continuously. Learners join for identity as much as capability. At that point, we are no longer an edtech startup. We are infrastructure for technical talent.

That is the ambition.

What Comes Next

Many operational questions naturally follow:

  • Detailed financial models
  • Legal structure
  • Program design mechanics
  • Technology infrastructure
  • Strategy sequencing

These are addressed systematically in the next chapter: FLOATS - Finance, Legal, Operations, Academics, Technology, Strategy.

This chapter explains the philosophy. FLOATS explains execution.

The last generation of edtech companies optimized for growth before trust. We are trying to optimize for trust before growth.

It is slower. It is harder. It is also far more durable. Our guiding principle for now is: Don’t die. Build credibility. Let compounding do the rest.