DeweyLearn Raises $5M Series A for AI Skills Assessment

DeweyLearn, a New York-based education technology company, raised a $5M Series A to scale multimodal AI for skills assessment across K-12 education, higher education, healthcare training, and workforce learning. The hard problem is not content delivery. It is determining whether someone can actually perform the skill they claim to understand.

Co-founded by Luyen Chou, CEO, and Dirk Liebich, CTO, DeweyLearn publicly launched in 2025 after more than a year in stealth. The company is building AI that analyzes classroom audio, video, learning data, and institution-specific assessment standards to produce expert-level feedback at scale while keeping educators responsible for coaching, judgment, and final decisions. That matters because education is moving from standardized measurement toward demonstrated competency, and the institutions driving that shift need evaluation systems that can scale without reducing learning to another checkbox.

What Happened

DeweyLearn announced its $5M Series A in mid-July 2026, with public reporting appearing on July 15 and July 16. The round was led by SJF Ventures, with participation from Catalysis Capital, Morningside, Owl Ventures, and additional investors. The financing gives the company capital to continue product development and expand adoption across schools, healthcare training programs, employers, and workforce development organizations.

The company built its platform around a straightforward but persistent question: can technology help experts evaluate real-world performance without pretending multiple-choice tests are a proxy for mastery? DeweyLearn combines audio, video, contextual learning data, curriculum, and customer-specific standards to evaluate observable learner behavior. The objective is not to replace instructors, but to reduce repetitive evaluation work so educators can spend more time on coaching and the nuanced judgment that still requires human expertise.

Why This Matters

Education has never suffered from a shortage of information. It has suffered from a shortage of individualized feedback because one experienced instructor can recognize the difference between a competent student and an exceptional one, while scaling that judgment across thousands of learners consumes enormous faculty time. DeweyLearn is targeting that constraint.

The company's multimodal AI approach is important because competency looks different in every profession. A culinary instructor evaluates knife technique differently than a nursing educator evaluates clinical judgment, and a workforce trainer needs different performance signals than a university professor grading written work. DeweyLearn's architecture reflects that reality by adapting to institution-specific standards rather than relying on generic scoring models, making the platform more relevant to practical education, healthcare training, and workforce learning.

Market Context

Artificial intelligence continues reshaping education, but much of the market conversation still centers on content generation, tutoring assistants, and administrative automation. Assessment represents a different frontier because it is closer to the outcome institutions actually care about: whether a learner can apply knowledge in a real setting, under real constraints, with enough confidence to move forward.

DeweyLearn's customer and partner activity reinforces that opportunity. The company reports evaluating more than 20,000 student homework submissions at the Auguste Escoffier School of Culinary Arts, helping faculty save hundreds of hours of grading time while preserving instructor oversight. Publicly disclosed work with Riverside Insights and NeuroAffective Relational Model also suggests the platform is being tested across specialized learning environments where assessment quality carries meaningful operational importance.

Competitive Signals

Real adoption often tells a more useful story than polished product language. DeweyLearn's use across culinary education, clinical training, K-12 learning, higher education, and workforce development gives the company a broader perspective than a single education technology workflow. The common thread is performance-based assessment in environments where expert review is valuable, expensive, and difficult to scale.

Recognition has followed that traction. DeweyLearn won the 2026 ASU+GSV Cup, earning top honors from a field of more than 3,000 education technology companies. That does not guarantee market leadership, but it does suggest the company's thesis is resonating across the education technology ecosystem. Awards open doors, but sustained customer outcomes determine whether those doors remain open.

What This Signals

The investment provides a useful signal about where venture capital continues to support education technology. Investors are becoming less interested in generic AI messaging and more focused on companies demonstrating measurable operational outcomes, credible customer deployments, and well-defined workflows where AI improves speed, consistency, or quality. DeweyLearn fits that profile because assessment is both operationally difficult and budget-relevant for the institutions it serves.

For founders, the lesson is straightforward. Capital follows evidence, and DeweyLearn's story is built around a concrete problem, visible deployments, and a product designed to make expert feedback available at a scale education has struggled to deliver. The technology may be new, but the investment logic remains familiar: solve an expensive problem, prove customers care, and demonstrate that the market extends well beyond the initial use case.

The Bigger Industry Shift

The long-term opportunity extends beyond classrooms. Healthcare, manufacturing, skilled trades, enterprise training, and professional certification all depend on evaluating observable human performance, and each market faces a similar constraint: experts remain limited while demand for high-quality instruction continues to grow.

DeweyLearn's broader vision centers on building a learning model that understands skills across multiple domains while adapting to each customer's evaluation criteria. If that balance holds, AI may become less interesting as a machine that generates answers and more valuable as infrastructure for helping people demonstrate genuine competence. That is not the familiar debate about replacing teachers. It is a more useful question about whether expert feedback can finally become more accessible without diminishing what expertise actually means.