The Fifth Elephant
The Fifth Elephant

The Fifth Elephant 2026 Winter Edition

Call for Data Problems and Proposals

All submissions

Craig Dsouza

@craigds

Practice + Feedback -> Smarter Children ?

Submitted Oct 9, 2026

🎤 I am submitting: To speak I have a submission for: A talk - 30-40 mins

Submission type

  • Talk/session proposal - 30-40 mins OR
  • Showcase - new ideas - 20 mins

Bio
Craig Dsouza , Head of Product & Platforms at Open Links Foundation

Open Links Foundation (OLF) is a not for profit , edu-tech company, that has a mobile app with a large user base of administrators and teachers (~250,000) primarily in Maharashtra.

The app, Vinoba, provides value to teachers with a Social feed that only teachers post to. This amplifies teaching programmes and activities even in smaller, rural districts, motivating teachers to do better. In new experiments AI is being used to analyse the videos and give teachers feedback on individual students’ capacities.

It also provides teachers access to learning materials for different subjects.

Lastly it provides teachers and administrators a view into which students and schools are doing well and which ones are doing poorly, in practice exams that OLF conducts. The practice exams are an entirely offline process, where answer input is via OMR sheets and grading is automated.

The larger theory of change of OLF is

  1. practice exams &
  2. in app experiences
    lead to teachers that are more motivated and better informed and this leads to better academic and holistic outcomes for children.

Session Description
We want to understand causality for two different questions

  1. what actions by teachers in-app are followed by other actions
    For instance, do teachers who view AI feedback on their posts, actually post more and is there an increase in the quality of their posts over time?

  2. what actions in-app can be causally linked to improved academic outcomes.
    Example: Do schools where teachers begin to use Vinoba actively see measurably better academic outcomes in the next semesters?

If our current data/process is insufficient to arrive at these conclusions, what do we need to change to get us there? e.g. we don’t run A/B tests. Are these strictly necessary for causal inference?

We discussed these and iterated on a plan with Claude (PDF presentation shared). We now intend to follow through with some parts of the plan and present our findings at the Winter Conference.

Key Takeaways
To be determined

Which audiences will benefit most from your session?
Engineers interested in education & impact focused engineering and data science challenges.

Slides
https://drive.google.com/file/d/1uPy3Lz7ZJixyErQu1AL8MTr5Rr14rGql/view?usp=sharing

What don’t you know yet, or what do you want help with?
This can be an open question, unresolved challenge, missing dataset, technique you are unsure about, implementation blocker, or a place where you want critique.

We could use help from data science & other education domain experts to help review our process, outputs and sharpen our questions.

Match making tags - choose any that apply:

  • I can help with: ideas / experience / dataset / critique / facilitation
  • I need help with: experience / technique / critique
  • I’d like to meet: researchers / practitioners / domain experts / tool builders / dataset owners/ facilitators

Tags
data for good, causal inference, education

How do you want to help? Choose any that apply:

  • facilitate a problem exchange session

Availability: are you available before the event for light-touch matching/facilitation?
Yes, only on weekends.

All submissions

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

The Fifth Elephant
The Fifth Elephant
Jumpstart better data engineering and AI futures