Goti Mastery

Goti Mastery is an AI-powered training platform, starting with martial arts, that connects academies, athletes, and tournaments with Gemini-powered video assessment and coaching insights.

Comment

Inspiration

What happens when martial arts training reaches thousands more children—but the number of experienced coaches doesn't scale with them?

That question became especially relevant to us in Bangladesh.

In August 2026, Bangladesh announced plans to give martial arts greater importance in compulsory sports education for primary-level students, alongside efforts to develop practical plans for expanding training.

We are not representing or working on behalf of the government. But the announcement highlighted a problem we were already thinking about:

If participation grows, how do experienced coaches support many more students without sacrificing the human guidance that martial arts depends on?

A coach cannot be everywhere a student practices.

Students may train for hours between direct interactions with instructors. During that time, mistakes can be repeated without structured feedback. As training expands across schools, academies, and communities, that gap can become even larger.

We saw an opportunity for AI—not to replace coaches, but to extend their reach.

Imagine one experienced coach supporting several schools or studios.

The coach teaches in person. Between sessions, students record their practice. Gemini analyzes the training video and provides structured observations, helping students understand what to work on before their next lesson.

When they meet again, the coach has better visibility into what happened between sessions and remains responsible for progression, grading, and professional decisions.

That became Goti Mastery:

Train → Assess → Improve → Compete

We started with martial arts because the problem is immediate, human, and visible.

But the underlying question is much larger:

How can AI help scarce human expertise reach more people without replacing the expert?

What it does

Goti Mastery is an AI-powered training and tournament operating platform, starting with martial arts.

It connects the athlete's journey in one ecosystem:

Academy → Coach → Student → AI Training → Progress → Tournament

🥋 Train

Academies can manage multiple studios, coaches, students, guardians, ranks, attendance, and permissions.

The platform is multi-tenant and role-aware, giving academy owners, studio administrators, coaches, students, guardians, referees, and tournament staff experiences appropriate to their responsibilities.

✨ Assess with Gemini

Students upload real training videos.

Google Gemini analyzes the performance and transforms the video into structured training observations, including:

  • detected discipline and form
  • overall performance score
  • strengths
  • areas for improvement
  • actionable recommendations
  • detailed assessment feedback

Assessment output is available in English and Bangla.

That matters to us.

Goti Mastery is designed not only as a technical demonstration, but as a system that can become useful in the communities where we are beginning.

And we deliberately keep one boundary clear:

AI assists. The coach decides.

Gemini does not award belts or replace instructors. Coaches remain responsible for grading, progression, and professional decisions.

📈 Improve

An assessment does not disappear after the AI response.

It becomes part of the student's training history.

This changes the feedback loop from:

Practice → Wait → Coach

into:

Practice → AI Observation → Improve → Coach Review

The goal is simple: help students make the time between coaching sessions more productive while giving coaches better visibility into their progress.

🏆 Compete

Training eventually leads to competition.

Goti Mastery therefore includes tournament operations in the same platform:

  • tournament lifecycle management
  • rulesets and divisions
  • athlete eligibility
  • officials and teams
  • registration
  • payment tracking
  • weigh-in and medical clearance
  • mats
  • bracket generation
  • match and scoring foundations

This is also central to our business model.

Tournament operations are intended to become a primary revenue engine for Goti Mastery.

Academies can train athletes, prepare them, register them, and eventually compete through the same ecosystem.

The journey becomes:

Train → Assess → Improve → Compete → Learn → Train again.

How we built it

We built Goti Mastery as a real multi-tenant SaaS platform, not a standalone Gemini demo.

Technology

  • AI: Google Gemini
  • Cloud: Google Cloud Platform
  • Backend: Python + FastAPI
  • Frontend: Next.js + TypeScript
  • Database: PostgreSQL
  • Cache: Redis
  • Security: JWT authentication + permission-driven RBAC
  • Architecture: Organization and studio-level multi-tenancy

The production assessment flow is:

Student Video → Validation + Student Context → Gemini → Structured Assessment → Persistence → English/Bangla Feedback → Coach Review

Gemini therefore operates inside an actual student workflow rather than as a separate chatbot.

The same platform foundation connects training with competition:

Organization → Studio → Athlete → Assessment → Tournament → Division → Registration → Mat → Match

That shared domain is important to our vision: the athlete receiving AI-assisted feedback today can be the same athlete competing through the platform tomorrow.

Challenges we ran into

Where should AI stop?

This became our most important product decision.

An AI-generated score should not automatically promote a martial artist, award a belt, or overrule professional judgment.

We designed Gemini as an observation and assistance layer while keeping humans responsible for consequential decisions.

Turning AI video analysis into a product

Calling Gemini was only one step.

Real users meant handling video uploads, validation, storage, processing states, provider failures, usage controls, persistence, multilingual results, and understandable error handling.

Modeling the real world

Running academies and tournaments involves complicated relationships.

An academy owner, coach, student, guardian, referee, and volunteer cannot have the same permissions or see the same information.

Tournament eligibility can also depend on rank, age, gender, weight, registration state, payment, weigh-in, and medical clearance.

Building those real workflows was significantly harder than building a demo screen.

Accomplishments that we're proud of

Our biggest accomplishment is that Goti Mastery moved beyond the prototype stage.

Today:

  • 🌐 Goti Mastery is live on Google Cloud
  • 🥋 real academy workflows can run on the platform
  • 🎥 students can upload real martial-arts training footage
  • ✨ Gemini analyzes that footage in production
  • 🇧🇩 assessment feedback works in English and Bangla
  • 📈 results become part of the student's assessment history
  • 👨‍🏫 coaches remain inside the decision loop
  • 🏆 training and tournament operations exist in one ecosystem
  • 🤝 we have begun real-world academy and commercial validation

The achievement we're most proud of isn't one feature.

It's connecting them.

We didn't build an AI martial-arts demo.

We built the beginning of an operating system for martial-arts training and competition.

What we learned

The Gemini API call was not the hardest part.

The harder problem was everything required to make AI useful in the real world:

identity + permissions + domain context + persistence + multilingual feedback + usage controls + failure handling + human oversight

And that led to our most important lesson:

AI is most valuable here when it scales scarce human expertise—not when it tries to replace it.

A coach cannot watch every minute of every student's practice.

AI can help observe some of those missing moments.

The coach can then spend limited human attention where experience, judgment, motivation, and responsibility matter most.

What's next for Goti Mastery

🏆 Run real tournaments

Our immediate priority is to operate friendly and championship tournaments through Goti Mastery and validate tournament operations as a primary revenue channel.

We are also continuing to mature our payment infrastructure as the business grows.

🧠 Build longitudinal training intelligence

Today Gemini can help answer:

"How did I perform?"

Next, we want Goti Mastery to help answer:

"What improved over the last 90 days?"

"Which mistakes keep recurring?"

"What should I practice next?"

That transforms individual assessments into a longer-term training intelligence layer for students and coaches.

🥋 Connect competition back to training

Competition should not be the end of the data journey.

Tournament and match observations can eventually become additional signals that help coaches understand what an athlete should work on next.

AI can also assist tournament officials with relevant observations and information while human referees and judges remain responsible for official decisions.

🌍 Expand beyond martial arts

Martial arts is our starting point, not the architectural limit.

Goti Mastery is built around organizations, studios, instructors, students, progression, assessments, and events.

That gives us a path toward other instructor-led disciplines such as Taekwondo, Judo, Yoga, Dance, and other skill-based training environments.

But we are starting with the community and problem directly in front of us.

Our long-term vision is simple:

Make expert-guided training more continuous, measurable, and scalable—without removing the expert.

One coach. More students. More practice. Better visibility.

Train → Assess → Improve → Compete

Built With

Share this project:

Updates

Submission history