4,200 children. 4,000+ volunteers. 140+ learning centres across India. Our volunteers are incredible
humans who give 2.5 hours a week to teach a child. The one thing that would boost the learning
outcomes they provide access to? Preparation.
You'd build an AI Coach that helps volunteers prepare for class in 2 minutes on their phone. Lower the
barrier, volunteers prepare more, kids learn faster. That's the theory. You'd make it real.
The Problem
U&I has a curriculum we've built ourselves. Student data, session notes, curriculum content and volunteer
activity currently live in different places in different formats. Nothing talks to anything else. Volunteers
navigate multiple platforms to prepare for a single class, and most don't prepare at all.
We need someone who can turn this into a system that delivers ready lesson plans to volunteers on
WhatsApp before class. Simple enough that it works. Cheap enough that a nonprofit can run it at scale.
What You'd Build
• A data architecture that unifies multiple data sources into a coherent system usable by AI.
• A RAG pipeline that grounds LLM outputs in real curriculum content. No hallucination.
• A data quality layer that handles vague, incomplete, real-world data and knows what to trust.
• A mobile delivery interface via WhatsApp. Dead simple. Every extra tap is friction.
• A cloud infrastructure layer that keeps it reliable and cost-efficient across 4,000 volunteers.
• A cost model that holds at scale. In a nonprofit, cost per query matters as much as output quality.
You're a Fit If
• You can design data systems, not just write code. Figuring out how data should flow between platforms,
what the architecture looks like and how to make messy nonprofit data usable for AI is the core of this
job.
• You've worked with LLMs and retrieval-augmented generation. You know how to ground model outputs
in real data and keep them constrained.
• You can ship a working prototype in weeks, not months.
• You're comfortable with ambiguity. Requirements evolve as the pilot generates feedback.
• You can explain technical decisions to non-technical stakeholders.Nice to Have
Airtable experience. Indic language experience. WhatsApp API platforms and other mobile-first solutions.
Education sector familiarity. Open-source contributions. Understanding of data privacy with children. None of
these are dealbreakers.
The Details
Type: Full-time, contract-based (consultant). Duration: Initial contract through Jun 2026, with a possibility of
extension based on program success. Location: Remote/flexible, with visits to learning centres during pilot.
Compensation: Competitive and fair. We can't match FAANG, but we're not trying to underpay either.
U&I is a volunteer-driven charity. The largest volunteer tutoring movement in India. Great Place to Work
certified. Flexible, informal, direct.
🚀 Before You Hit "Apply Now"...
This form will take around 20 minutes.
Grab a drink, get comfy - no rush.
Why’s it long?
Because we want to know the fun, human side of you before we show up with our capes on.
