read part I first

Paulo Coelho de Souza, born in 1947 in Rio de Janeiro, once said,
“The world is changed by your example, not by your opinion.”
You might have heard of his little known 1988 novel The Alchemist, which somehow became a global best-seller. Beyond the books, through the Paulo Coelho Institute, he supports disadvantaged Brazilians and fights poverty.
This quote is special to me because it perfectly embodies what was revealed to me after my 2023 visit to the Malto tribe. At the time, I didn’t have a plan or a blueprint. I wasn’t thinking in terms of systems, software, or solutions. I was just there. Observing, listening, and trying to understand a world far removed from the assumptions I was used to. The experience felt heavy, but unfinished.
Months passed. Life moved on — school, projects, deadlines. Yet Jharkhand never fully left my thoughts. What eventually changed was not the memory itself, but the realization that remembering is passive. My opinion alone cannot change outcomes. It only creates a choice about what to do next. That choice became the spark for the work I would eventually start.
The Nudge
The push didn’t come from research papers or policy debates. It came from a conversation with my father. Over time, he reminded me of that 2023 trip and gently pointed out something I had been circling but not fully acknowledged. I had a unique perspective, a close-to-home experience, and skills that could actually make a difference. He restated the problem, not as a lecture, but as a question that lingered in my mind.

I wasn’t suddenly discovering injustice or waking up to inequality. I had already seen it. What changed was the understanding that knowing something and not acting is still a choice, and sometimes a costly one. Awareness alone, no matter how vivid or painful, does nothing for the people on the ground. The difference comes from taking even small steps with the resources you have. That realization made the problem tangible. It made the abstract concrete. It turned observation into responsibility.
I didn’t need to become an educator, policymaker, or activist. I didn’t need to try to solve education as a whole. What I could do and what I needed to do was build something small, careful, and practical. Something that worked with the people already doing the work, respected the communities and their culture, and addressed a specific gap in a meaningful way. That was the moment the idea stopped being abstract and started becoming a plan.
Why Attendance Came First

When I thought about where to start, it became clear that the most immediate and actionable problem was simply knowing who was actually in school. Attendance might seem basic, almost trivial, but in the Malto context it was a huge barrier. Students were coming and going unpredictably. Some walked hours along dangerous paths, others missed school for chores or illness. Teachers struggled to track who was present and who was quietly falling behind. Without reliable attendance data, identifying students at risk of dropping out was almost impossible.
It was not about motivation or discipline. Students wanted to learn. They valued school. The system made consistent attendance fragile. From my time in Jharkhand and conversations with missionaries and school staff, one thing was clear: schooling itself was unstable. Attendance depended on weather, terrain, health, family responsibilities, and distance. Systems that looked functional on paper often failed in practice.

The bigger problem was invisibility. If a student stopped coming, there was no reliable way to notice early. Records were scattered or incomplete. Absence became a number, not a signal. That gap mattered because you cannot respond to what you cannot see.
I want to be clear: technology does not fix education inequality. It never has. An attendance app can easily turn into surveillance. It can become a tool of pressure. It can unintentionally push western ideas of productivity, discipline, and success onto communities that operate on different values.

This concern came up repeatedly. People who had worked with the Malto community for decades told me directly:
We do not want to become the police.
We do not want to force kids into school at the cost of their culture.
We do not want education to erase identity.
The goal was never enforcement. It was visibility without coercion. A system that helps educators notice patterns and ask better questions, not punish outcomes.
Building With Constraints, Not Against Them
Once I committed to building something, the constraints became the design.
Reliable internet and electricity are rare. Devices are mostly low-end Android phones, often under $100. Teachers may only get connectivity once a day. Any system built like a typical online app would fail instantly.

Enter rolecaller. An offline-first, local-first platform built for classrooms where the internet can’t be trusted. This isn’t a standard CRUD app. At its core, rolecaller is a distributed system designed to survive broken networks — where the teacher’s device is the source of truth, and the cloud exists only as a convergence point for data.
rolecaller has two main views: Admin and Teacher. Admin view lets school staff see overall attendance for the entire school or for individual students. Admins can create classes, add students, and assign students to specific classes.



Teacher view is used for day-to-day attendance. Teachers log in, select a class, and record attendance. Every teacher has access to all classes. Local-First by Design
I’m now going into technicalities (if you want to skip this part that’s okay)
Each device runs its own SQLite database using Expo SQLite. Attendance is recorded locally, instantly, without waiting on the cloud. The cloud exists only as a convergence point, not a dependency.

This matters because it respects the reality of the environment instead of fighting it.
The Pulse Sync Engine
The heart of the system is a custom synchronization layer called the Pulse Sync Engine.

Its job is simple in theory but hard in practice. It safely moves data between the teacher’s device (the edge) and the NEON cloud database whenever the network allows, without losing data, duplicating records, or blocking the user.
The sync process follows a strict push and pull protocol.
Push (Upstream Sync):
- The app watches for stable connectivity.
- It looks for attendance records marked as unsynced in the local SQLite database.
- Records are batched intelligently to reduce payload size.
- Data is only marked as synced after the server confirms success.
- The backend uses idempotent UPSERT operations, so duplicate packets don’t corrupt data.
If the app crashes mid-sync or the signal drops, nothing breaks. The data stays local until it’s safely confirmed.
Pull (Downstream Sync):
- When connectivity is verified, the app fetches updated school data from the server.
- It performs a “smart rehydration,” refreshing rosters and class lists while preserving unsynced local attendance.
- Teachers always see up-to-date information without risking the loss of their day’s work.
This design choice wasn’t about being clever, instead, it was about not punishing users for bad infrastructure.

Why UUIDs Matter More Than They Sound
Every entity in the system, schools, classes, students, attendance records, uses UUID v4 identifiers. That decision solves a major problem in offline systems: distributed ID generation.

Because devices can generate valid IDs offline, they don’t need to ask the server for permission to create records. When data finally syncs, everything fits cleanly into the central database without collisions.

It’s a small detail that makes offline autonomy possible.
Security Without Locking People Out
Authentication had to work in two realities: connected and disconnected.

When the device is online, teachers log in normally through the system. Login details are securely stored, and a session is saved on the device. If the internet goes down, the app switches to offline mode. Teachers can still log in using saved credentials and take attendance without a connection.
Blocking access during outages would defeat the purpose of the system.
Why the Backend Is Intentionally Simple
The API layer was deliberately built without heavy frameworks. Using raw Node.js HTTP and a shared PostgreSQL connection pool helps minimize cold starts and reduce overhead, while still handling bursts of sync requests from multiple devices during the short windows of connectivity that the schools experience. Drizzle ORM was selected not for abstraction, but for efficiency as it generates raw SQL with zero runtime overhead, ensuring that sync operations remain fast and predictable. Every choice in the stack was guided by the realities of the environment rather than convenience.
If you want to know more about how I built this app or more about the stack please make sure to check out the GitHub! please star it too🥹
Testing Before Trust
In the weeks leading up to my January 26 visit, the app was tested in collaboration with educators and missionaries. We observed how it performed under real conditions, iterating on flows, removing unnecessary features, and simplifying the interface to make it as intuitive as possible. Rather than going through the Google Play Store, we decided to distribute the app as an APK file to avoid bureaucracy and keep it simple for the schools. When I finally arrived in the hills, I spent a lot of time walking people through how the app worked. I showed it to teachers, administrators, and missionaries, and we had extensive conversations about its implications, potential pitfalls, and other challenges on the ground, including attendance barriers, cultural considerations, and maintaining student autonomy. After those discussions and final refinements, rolecaller was presented to the main board of the FMPB organization last week and approved for rollout.


The plan is to start small — beginning with RCPSC and a few smaller schools — before expanding to larger learning centers. This isn’t about scaling fast. It’s about learning carefully.
What Comes Next

Future updates will likely include a consolidated panel to view trends across schools, but only once enough meaningful data exists to make it useful. Expansion will be gradual and responsive, guided by the needs of the community and educators on the ground rather than imposed from above.

Education here is not about producing standardized outcomes. It is about sustaining opportunity and supporting learning in a way that respects and preserves the Malto community’s culture.
Closing Thoughts
This project did not come from a sudden burst of empathy or a desire to “fix” something. It grew from sitting with a memory long enough to understand that responsibility does not always arrive urgently, but it does arrive. It came from realizing that my skills and experiences could be applied in a practical, respectful, and meaningful way.

This app is not a solution in itself. It is a support structure. A quiet one. Its success will be measured not in visibility or recognition, but in the small, consistent ways it makes the work of teachers and communities easier and more connected. Sometimes impact is not about adding complexity. It is about removing friction and giving opportunity the space to grow.

It started as a simple opinion, which grew into a question, then an idea, and now, I hope it can become an example of how technology can support education, even in the most remote corners of India. Thank you for taking the time to read about this journey. I hope sharing my experiences in Jharkhand and the work with the Malto community gives you a clearer sense of the people, the challenges they face, and the impact that thoughtful, careful action can create. God Bless🙏.