In this episode of Chalk and Code, host Theron Davis sits down with Raj Valli, Founder and CEO of Thinkster Math, an AI-driven tutoring company that blends patented AI with human coaches to deliver measurable outcomes in math learning.
Drawing on experience across 12 industries and decades of nonprofit work in education, Raj unpacks why so many students still struggle with foundational math and what it really takes to design systems that reliably close learning gaps instead of just delivering content.

From AI-powered diagnostics that probe “wells of knowledge” to deep analytics that surface patterns like “this student struggles with two-digit multiplication when numbers end in 9,” Raj explains how Thinkster uses data, instructional design, and human tutors together to guarantee outcomes rather than hope for them. He also explores the limits of today’s AI, arguing that most models still rely on crystallized intelligence, while truly transformative learning and innovation depend on human fluid intelligence.

If you care about how AI, behavior science, and human relationships can be combined to help learners “learn better, faster, sooner,” this conversation offers a practical, grounded look at what high-accountability edtech can be

Key Moments

00:00 Welcome to Chalk and Code and introduction to the episode

01:22 Raj’s journey across 12 industries and nonprofit education work in India

02:21 The “airlines” analogy and the shocking reality that 40–50% of students are failing in math basics

04:14 Why “passionately curious” is the mindset behind Thinkster’s constant iteration

07:08 Positioning Thinkster as a learner-centric, “cradle to grave” solution rather than just a K–12 tool

09:55 The five key nodes: AI, data and analytics, curriculum design, human-in-the-loop tutoring, and behavior

13:48 Why behavior, discipline, and intrinsic motivation are as critical as content and technology

16:10 Rethinking diagnostics: using AI “dipsticks” to find where a learner’s true starting point should be

19:12 How probabilistic assessment works and why diagnostics must feel confidence-building, not punitive

23:24 Designing for emotion as well as intelligence so learners feel more confident as they progress

27:08 Giving tutors real insight, not just data: “this student struggles with numbers ending in 9”

29:44 AI as “actual intelligence”: why human judgment is still essential for context and bias

31:49 Crystallized vs. fluid intelligence and why Raj is skeptical of claims that AI will fully replace humans

35:31 Shifting from delivery of knowledge to adaptive use of knowledge as the core goal of education

36:40 Why adaptive knowledge and real-world application matter more than memorization

Why You’ll Love This Episode

This episode challenges the default assumptions behind most learning technology. Instead of starting with “What can AI do?”, Raj starts with “What does the learner need, right now, to move forward?” and builds from there.

You’ll hear how Thinkster:

• Uses AI diagnostics and behavioral data to find each learner’s true starting point and close gaps with precision.
• Combines human tutors, curriculum design, analytics, and behavior science into a single, outcome-focused model.
• Gives tutors actionable insights instead of dashboards for their own sake, so every session is targeted and high impact.
• Frames AI as a powerful tool for crystallized intelligence while insisting that fluid intelligence, context, and values remain deeply human responsibilities.
• Re-centers education around adaptive knowledge: not just what students can recall, but what they can do with what they know.

Connect with Raj Valli
LinkedIn: https://www.linkedin.com/in/rajvalli
Thinkster Math: https://hellothinkster.com/

Connect with Us
Company’s LinkedIn: https://www.linkedin.com/company/flexion-inc
Company Website: https://flexion.us/

If this conversation sparked a new idea or challenged your assumptions, share this episode with a colleague and subscribe for more thoughtful discussions at the intersection of learning, leadership, and educational technology.

Join us next time on Chalk and Code, where the science of learning meets modern technology.