THE CHALLENGE: The Phase II “Production Gap”

Securing $1M in federal Phase II funding marks a critical shift from “proving a concept” to “building a scalable business.” For this K–8 mathematics initiative, the stakes were high: the $1M grant depended on the team’s ability to take their theoretical learning model and transform it into a production-ready platform capable of supporting nationwide classroom pilots.

The transition introduced three major engineering hurdles:

  1. Translating Pedagogy to UX: Moving a complex model into an interface where students could independently question mathematical scenarios without constant teacher intervention.
  2. Multimodal Expression: Developing tools that allowed students to visualize abstract principles through a combination of text and digital drawing.

The Reliability Mandate: With large-scale classroom pilots scheduled, “downtime” or “regressions” weren’t just technical glitches—they were direct threats to research validity and future funding.

THE STRATEGY: A Lifecycle Partnership (Phase I to Phase II)

Unlike traditional agencies, we acted as the project’s technical co-founder. Because we built the initial prototype that successfully won the Phase I grant, we were uniquely positioned to architect a Phase II ecosystem that balanced educational fidelity with commercial-grade stability.

1. The AI-Powered Multimodal Sandbox

We engineered a sophisticated “storytelling engine” that empowers students to move beyond passive consumption.

  • Dynamic Scaffolding: We built interactive workflows that guide students as they question the core math principles behind a story, ensuring they develop foundational understanding before moving to expression.
  • The Visualization Layer: We designed a custom drawing and annotation tool where students visualize their unique mathematical models. This interface captures student inputs and orchestrates a hybrid AI process—hosting the frontend on React/Firebase while leveraging a specialized AWS-hosted AI engine to generate narrative-driven responses and custom imagery based on the student’s “storyboard.”
  • Creative Synthesis: The final stage of the workflow allows students to create their own narrative-based math problems. Our logic allows the AI to dynamically adjust problem difficulty based on the student’s grade level, ensuring the platform remains challenging yet accessible for the entire 5th–8th grade spectrum.

2. Research-Grade Data Infrastructure

Research is only as good as the data it produces. We built a FERPA-compliant data logging infrastructure designed for the rigors of high-level academic scrutiny.

  • Researcher Insights: We developed summary reporting tools that allow Principal Investigators (PIs) to pull clean, structured data for analysis.
  • Collaborative Interoperability: The system was built to allow renowned 3rd-party research partners to perform independent efficacy studies without compromising system security or student privacy.

3. Enterprise DevOps for Grant Milestones

To support the $1M funding roadmap, we moved the project into a production-grade delivery cycle. This included automated end-to-end testing and stable deployment workflows, ensuring that classroom pilots across multiple districts were never interrupted by technical regressions.

THE IMPACT: A Commercial-Ready Research Asset

By serving as the core engineering team from day one, we helped the project achieve “Commercial Grade” status while maintaining “Research Integrity”:

  • Secured $1M in Funding: The initial prototype we developed was instrumental in the successful Phase II grant application.
  • Active Independent Research: The platform now supports high-fidelity data collection for ongoing studies with global research organizations.
  • Scaled AI Integration: Successfully balanced a hybrid cloud architecture (Firebase + AWS) to provide a smooth, low-latency experience for student creators.
  • Empowered Student Autonomy: The platform successfully guides students as they question, visualize, and create, fostering a sense of ownership over their mathematical learning.

THE VERDICT

“Phase II funding demands a transition from ‘research project’ to ‘software company.’ We provided the engineering discipline and product vision to ensure the technology didn’t just support the research—it accelerated the mission.”