AI-augmented PBL Authoring Toolkit

The Problem & Opportunity:

As AI reshapes our world, parents and educators worry that traditional schooling won't prepare students for an ever-changing future that demands durable skills.

I run a growing project-based learning (PBL) community. Every day, educators and school leaders reach out for help building PBL curriculum—even for kindergarteners. They ask the same questions: How do we get started? How do we align with standards without becoming overly rigid or sacrificing creativity? How do we assess learning meaningfully? How can we do this with limited time and resources?

Despite my eagerness to help, I can't do it alone. Even for experienced learning designers, turning one promising idea (or set of standards) into a coherent, standards-aligned project can take weeks. Without care and rigor, projects drift into the wide-but-shallow middle: "themed activities" that check the box but lack the genuine questions,  authentic audiences, clear learning goals, rigorous progression, and meaningful assessments that make PBL transformative.

The core problem is clear: interest is high, but authentic and quality PBL is hard to scale. How might we empower educators, especially ones with limited experience and resources, to design rigorous PBL learning experiences at scale?

Product Vision & Strategy:

We make authentic, rigorous PBL repeatable at scale by scaffolding the hardest design decisions : 

• Setting clear learning goals (including interdisciplinary connections)

• Sequencing lessons so concepts deepen and connect

• Finding and adapting vetted resources

• Assessing learning with confidence —all without requiring design expertise or extensive time and resources.

PBL cannot be one-size-fits-all, but the design process can be scaffolded. We take an opinionated but composable approach and aim to build a toolkit grounded in learning science. Each tool comes with examples and templates, while the system stays flexible so educators can adapt components to fit their context.

Key features include:

• Knowledge Map–Driven Curriculum Authoring: Educators input standards and ideas to generate a knowledge map, which they can then refine to plan a learning progression that evolves flexibly with student inquiry.

• Resource Search (aligned to the map): The tool recommends high-quality, evidence-based materials aligned to specific knowledge components.

• Assessment Management + Learning Analytics: Educators collect student work and responses tied to knowledge components, with student- and class-level analytics to support dialogue, feedback, and iteration.

The following are screenshots of our early prototypes:

Current Traction & Roadmap:

We are executing a phased rollout strategy to evolve the toolkit from a high-touch prototype to a scalable platform.

We begin by partnering with two pilot sites to co-create experimental modules (Kindergarten STEM & 5th-grade AI), using these sessions to manually validate the AI workflow and establish baseline metrics. 

Subsequently, we will deploy the Alpha version to 2–4 partners. We will provide high-touch onboarding and ongoing scaffolding to ensure these partners can successfully utilize the platform for knowledge mapping and assessment design. This hands-on approach allows us to closely monitor usage friction and iterate on the user experience before scaling to 15–20 organizations in the Beta phase

Learning Science & Design:

Our approach is grounded in these key frameworks:

• Learning Engineering (Knowledge–Learning–Instruction Framework): We treat learning as measurable Knowledge Components (KCs) and use the KLI dependency principle to recommend instructional strategies tailored to the specific type of knowledge.

• Emergent Design: We support an iterative cycle where goals and content evolve based on classroom evidence, rather than a rigid linear sequence.

• Concept Mapping: Integrated into the workflow to visualize prior knowledge and support collaborative knowledge construction.

Design Process:

User Interviews:

I recruited education leaders, teachers, and parents with curriculum development needs from my PBL community. I led the user research interviews and collaborated with my partner to synthesize findings into user personas.

Design with Users:

To ensure the toolkit solves real-world problems, we prioritize deep collaboration with early users ("do things that don't scale"). We operate as research partners: co-designing pilot modules alongside teachers to stress-test the AI workflow in actual classrooms.

Driven by organic inbound demand, we have 5 confirmed partners across the U.S. and China, reaching 970+ students in 2026:

Pipeline & Strategy: We prioritize "market pull"—all current partners reached out organically to solve design bottlenecks. We are validating the toolkit across diverse contexts (public schools & after-school, US & China) and will expand recruitment via community referrals and structured open calls.

Prototyping:

I prototyped a teacher's dream flow of designing and implementing PBL experience.

I used a concept mapping tool for teachers (Lynkage) to prototype the process of building knowledge map.

I also used AI tools (ChatGPT and VS Code) to code early prototypes and iterated on it with my partner.

Benchmarking / Competitor Analysis:

I researched current PBL curriculum building and delivery tools, like Inkwire, PBLworks Teach, and Prenda School. 

I also draw inspiration from HCI labs like DynamicLand, research tool like ResearchRabbit and notetaking tool like Granola.

Learning Science Research

I developed product concept from research in PBL & formative assessment, concept mapping, knowledge construction & documentation tools.

Our Mission and What We Value

Our mission is to make learning design accessible to every educator and learner—providing practical tools and guidance to co-design, adapt, and improve learning experiences. This mission is rooted in three beliefs:

1. Education is the key to unlock each person’s full potential and solve society's systemic crises.

2. Agency is paramount. Everyone should have the power and tools to design education for themselves and their communities as their needs evolve.

3. Learning is a living, complex system, not a linear manufacturing process. Effective education must emerge from local realities rather than be imposed from above.

Education today fails to realize this potential because of:

1. Rigidity: Current educational pathways are linear and inflexible, driven by top-down curriculum design. With high opportunity costs to change, learners are often forced to disengage or drop out.

2. Exclusivity: High-quality learning design is not a capability accessible to the many. Only a few experts are trained to design education incorporating best practices.

We are dedicated to inventing a humane, AI-augmented authoring environment that supports everyone to design education that evolves with their realities. Our values are deeply inspired by Dynamicland, a non-profit research group from Berkeley reimagining computers as a medium for new thoughts. We use technology not to automate, but to amplify human creativity. 

We choose to start with project-based learning because PBL is a scalable pedagogical framework that explicitly mandates “authentic experiences" and community connection as a design requirement, making it the natural vehicle for contextual, emergent learning.

We begin by serving the group that knows education best—teachers and education leaders—empowering them to reclaim their role as designers and co-design with their students. From there, we will expand to parents and independent adult learners, and ultimately, put these learning design tools directly into the hands of youth and children.

We envision a future where education transcends the walls of classroom and training programs. Education is not a product we consume, but a lifelong continuous journey where every learner—child or adult—is empowered to design their own growth.