Why we made it a habit, not a gesture
At Protagona, we've learned that the value of our work isn't measured only in systems shipped or workloads optimized, but in who those systems ultimately serve. Working with nonprofits has a way of making that truth impossible to ignore.
When a dashboard built by our team helps a case worker reach a family faster, or a data pipeline means more meals get delivered before they spoil, engineering takes on a different dimension that's hard to find elsewhere. The standups feel different. The retros feel different. Even the bug tickets feel different, because the cost of a delay is no longer measured in a missed sprint goal. It's measured in someone who didn't get reached this week.
Many Protagonists are personally philanthropic or involved in their communities, and as we continue to grow we wanted to find a way to serve these organizations. In 2024, we started exploring what it would look like to give back as an organization in a sustainable way, and explored a few options: pro bono bench work, discounts, funding programs, and beyond. After collaborating directly with individuals on the front lines of the nonprofit sector, we recognized their primary requirement was to have a dependable partner capable of delivering results and maintaining a dedicated, long-term commitment to their organizational success.
The selfish version of why this matters: these projects are often the most technically challenging custom builds our engineers take on in a given year. Constraints sharpen design. A nonprofit's budget is unlike a traditional commercial organization in many ways, and often it's not a question of how much to spend but how to stretch every dollar for maximum impact. This generally means we have to make more deliberate architectural calls, lean on serverless and managed primitives more thoughtfully, and write documentation that actually gets read.
The less selfish version: it changes the kind of company we are. We provide a consistent discount to all nonprofit customers and prep our delivery teams to understand and accommodate how unique nonprofits are, both operationally and in their missions.
Three projects that shaped how we think about impact
1. A Food Bank Technology NPO: Redshift-backed analytics, 40% faster routing
A food bank serving roughly 8,000 pantries across the US was on a mission to reduce 11 billion pounds of food waste per year. Their existing routing process lived in a mix of spreadsheets and a legacy logistics tool that couldn't account for perishable-product time windows in any sane way.
We built a Redshift-backed analytics platform that consolidated their inventory, partner intake, and vehicle telemetry into a single warehouse, and layered on a routing-optimization view that surfaced the cost of each route in terms of both miles and time-to-spoilage. The headline number was a 40% reduction in routing time, but the operational impact mattered more: more deliveries hit the window before perishables turned, which meant less food written off and more meals reaching families.
What made it technically interesting was the data quality work. Half the value came from reconciling four years of partial records before we could trust any model on top of them. That's the unglamorous half of analytics work, and it was a useful reminder of how much production-grade engineering on a small nonprofit project actually parallels the work we do at scale for commercial clients.
2. Empowering Youth Services Through AI-Driven Innovation
A youth services nonprofit was in dire need to replace a rigid, outdated EHR that couldn't support their complex programs and compliance demands. We designed a modern, AI-forward platform that streamlines documentation and scheduling, with a long-term vision of becoming a white-labeled SaaS product.
We leaned heavily on Amazon Bedrock to anchor the solution in AI-driven utility and Smart Scheduling, which optimizes logistics based on real-time availability. The underlying architecture is modular and serverless, favoring an API-first design that ensures the system can scale without becoming a maintenance burden.
Our process and deliverables were strictly design-oriented, but with a foundation of testing and validation; we used clickable prototypes and rigorous system documentation to surface and resolve design flaws early. We optimized for the long game by designing a SaaS-ready foundation, allowing for future white-labeling and transforming a bespoke tool into a scalable asset that can serve the broader nonprofit ecosystem.
3. A Youth Literacy Nonprofit: Meaningful Reading at Scale
A nonprofit educational platform serving over 800,000 educators and 11 million students annually was seeking to revolutionize access to affordable, diverse level-appropriate reading materials in K-12 education. Addressing a major void in the education market, the organization focuses on low-income schools, where research indicates differentiated instruction is least common.
We developed a sophisticated generative AI system to scale the production of high-quality educational content. The AI model generates accurate and age-appropriate leveled texts, example written answers for question sets, and enhanced metadata to improve search functionality. The system adapts tone and complexity based on grade level, subject matter, and learning objectives while maintaining the client's rigorous editorial standards.
What's hard, honestly
Maintainability. The word can mean a lot of things, but in essence it's the ease with which a system can continue to function and provide its intended value after those who built it have moved on. The most elegant architecture in the world is a liability if the organization on the other end can't keep it alive.
I've been guilty of this in the past, and have learned the dangers of over-engineering in the pursuit of perfection. It took me a while to internalize that completeness isn't the goal. Fit is the goal. Not every organization has a team of senior data engineers on call. If we hand them something that requires deep AWS expertise to debug at 2am, we haven't solved a problem, we've just relocated it.
So we've gotten more disciplined. We default to managed services with predictable cost curves. We write documentation we'd actually want to read, with runbooks, decision logs, and what-to-do-when-X-breaks guides aimed at the personnel inheriting the system. The best measure of one of these engagements isn't what's running the day we leave. It's what's still running and still valuable, a year later.
Why this matters to me, in my role
I lead Engineering at Protagona. The honest answer to why I push for this work isn't only that it's good for the firm, though it is. It's that it's the part of the job that most reliably reminds me what we are here for.
Engineering is a craft that can drift into abstraction quickly. The systems get bigger, the dashboards get prettier, the architecture diagrams get cleaner, and somewhere in there, it's easy to forget that none of it matters unless someone on the other end is better off. Nonprofit work doesn't let you forget. It puts the people at the end of the pipeline in front of you, every standup.
If you're a technical leader weighing how to build social impact into your practice, my advice is to make it structural, not episodic. Reserve the capacity. Scope it like real work. And measure success a year out, not a month out.
Ethan Schumann
VP, Engineering | Protagona