Concept prototype by Anchit Tandon, with made-up students · not an official PW product
Learner Retention EngineBack to the PW case study
Interactive concept I built · Senior Product Manager - Growth

See what happens when a student starts to drift.

What I built: four fictional learner journeys, adjustable risk signals, a simulated catch-up experience, report previews and an editable financial model. Pick a learner and change the signals to see the concept respond. What this does not do: run trained AI, call PW APIs, use real learner data, contact parents or prove growth uplift. The rules and responses are illustrative.

Part of Anchit Tandon's Senior Product Manager - Growth interview case study for PW. Step-by-step review: anchit-tandon.com/pw-study#walkthrough

Concept prototype by Anchit Tandon · synthetic data · not affiliated with PhysicsWallah

On this screen: Drop-off Risk Score (engine output; the phone shows it as an on-track score), Catch-up Plan (the plan in the phone) and Next-Step Recommendation (bottom of the phone).

1 · Pick a student

Arjuna JEE 2028

Namaste Riya

On-track score 34

This week's on-track score

Needs a push this week

6 lectures behind batch pace

38%Lectures done
45%Test avg
0Day streak

AI Mentor plan · added to your Tracker 360 schedule

    Ask your AI Mentor Simulated

    Your next step

    2 · Drag the signals

    The phone and the score react

    Drop-off Risk Score: engine output

    Drop-off risk
    / 100

    This is a transparent rule-based illustration, not a trained ML model. A production system would first validate internal data access and quality, legal eligibility, and the performance of simple rules. Predictive ML would be added only if temporal calibration, precision and intervention experiments justify it.