Math equity for the next generation.

VietAlpha is the first math AI infrastructure built for the Vietnamese curriculum. High-quality tutoring should be accessible to every student, not just a privilege for those who can afford it.

For a rural family, one subject of one-on-one exam-prep tutoring can cost up to 160% of the average monthly income per person.

Read Our Research See the solution

The Problem

A paywall on the playbook.

Vietnam's math education is intensely exam-driven: success on national entrance exams hinges on material that reaches beyond the standard classroom, and that gap is typically filled by expensive private tutoring — putting advanced preparation out of reach for many.

Existing AI tools for Vietnamese students are often adapted from foreign systems. VietAlpha is built from the ground up to align with local standards.

Tutoring cost vs. income per person

Rural Household

Income per person · ~4.5M VND / mo

Group

~0.5M · 11%

1-on-1

up to ~7.2M

+60%

Premium 1-on-1 exam prep: up to ~7.2M VND/mo for one subject — 160% of income per person

National Average

Income per person · ~5.4M VND / mo

Group

~0.5M · 9%

1-on-1

up to ~7.2M

+33%

Group class ~0.5M · 1-on-1 exam prep up to ~7.2M VND/mo, per subject

Curriculum-aligned math datasets before VietAlpha

Income figures: GSO Household Living Standards Survey, 2024 (average income per person/month). Tutoring rates: HCMC & Hanoi tutoring-center price tables, 2024–25; premium exam-prep ≈ 450–600K VND/session. The wealthiest families spend up to 10× more than the poorest on out-of-school tutoring (OECD, 2025).

The Solution

Vietnam's first math AI infrastructure, built from the ground up.

VietAlpha is a reasoning dataset and AI ecosystem aligned with MoET standards. It provides pretrain, fine-tune, and evaluation splits for researchers and developers.

Curriculum Alignment

Built on official MoET textbooks and the national exam question bank.

Reasoning-First

A Socratic approach that encourages students to reason through problems rather than memorizing answers.

Continuous Practice

Generates drills across all difficulty levels, including exam-style questions.

Open Problem Library

A free resource powered by VietAlpha AI to provide guided problem-solving for every student.

How It Works

From a student's question to a complete answer in four steps.

Built on Vietnam's curriculum

Official MoET textbooks, real exam questions, and national exam formats are the foundation. Nothing foreign, nothing guessed.

Filters for accuracy

An accuracy layer removes confusing or incorrect content before it reaches the student. Quality is enforced at the source.

Matches the right lesson

Every question is matched to the specific topic the student is working on, keeping practice relevant and on-track.

Coaches, step by step

VietAlpha gives hints first, then guidance. Students arrive at answers by reasoning, not by being handed them.

Milestones

What VietAlpha has already achieved.

ICIA National Award

1st

National Youth Innovation Award, Edtech Category (UEH)

Recognized as a leading student-led edtech innovation in Vietnam

Dataset

1st

Curriculum-aligned math dataset built to Vietnam's MoET standards

Partners

Institutional and academic partners across Vietnam and the US

Pre-Seed

$4,500+

Raised in pre-seed funding to build Vietnam's math AI infrastructure

Infrastructure

Cloud support from AWS and Microsoft Azure for compute and AI services

Roadmap

Where VietAlpha is headed.

Feb 2026

Data Scaling

Accelerating the data pipeline and model training process.

Jul 2026

VietAlpha V1.0

Release of the first prototype for student testing.

Nov 2026

School Integration

Rollout to partner schools with direct feedback loops.

Mar 2027

Public Release

Open access for all students with continuous iteration.

Jun 2027

Expansion

Deepening integration across Vietnam's public school system.

Program Context

VietAlpha inside the CS Inspiration ecosystem.

01: Student Layer

Maths4Threads

45,000+ learners · 75 countries

Platform linking math practice to clothing donations for those in need.

02: Classroom Layer

Numbers to Networks

Vietnam · Indonesia · Malaysia

Inquiry-based math delivered in schools in partnership with regional Departments of Education.

03: Research Layer

VietAlpha

AI · Dataset · Infrastructure

The core layer providing the infrastructure for localized math AI tools.

04: Alumni Layer

Impact Fund

$75,000 capital

Student-managed fund supporting scholarships and seed funding for student ventures.

FAQ

Frequently asked questions

What is VietAlpha?

VietAlpha is a curriculum-aligned math AI and reasoning dataset built to Vietnam's Ministry of Education and Training (MoET) standards. It provides pretrain, fine-tune, and evaluation splits so researchers and developers can build math-reasoning models that follow the Vietnamese curriculum.

Who is VietAlpha for?

VietAlpha serves researchers and developers building Vietnamese math AI, and ultimately students who need affordable, curriculum-aligned math practice. The dataset and tooling are oriented around the classroom and the national-exam context.

How is VietAlpha aligned with the Vietnamese curriculum?

Its data is built from official MoET textbooks and the national exam question bank, with an accuracy-filtering stage and topic-level lesson mapping so that practice stays relevant to what students are actually studying.

Is VietAlpha free?

VietAlpha's open problem library is designed as a free resource for students, powered by the underlying dataset and models.

Who builds VietAlpha?

VietAlpha is a research program of CS Inspiration, conducted in association with Deerfield Academy and led by founder Quang (Harry) Tran. It receives cloud support from AWS and Microsoft Azure.

What is the current status of VietAlpha?

As of 2026, VietAlpha is scaling its data pipeline ahead of a V1.0 student-testing prototype planned for mid-2026, followed by integration with partner schools.

Continue reading from VietAlpha Research