Home / Our work / Quant Kid Academy

Live at quantkidacademy.comEdTechFinTechPrivate AI

Quant Kid Academy: a flight simulator for trading decisions.

A training platform where learners practise on real market history with virtual money and are scored on how well they decide, not on luck. A guided 8-week program, gamification and a private AI assistant take learners from first trade to running their own experiments.

IndustryEducation · Finance

ScopeProduct · Backend · Web · AI · DevOps

AudienceAdult learners & institutions

StatusLive in production

Visit quantkidacademy.com ↗

5,300+Automated tests: unit, UI and end-to-end

160API endpoints

40+App screens across 36 feature modules

100%Repeatable: the same inputs always give the same replay

The challenge

People learn trading by losing real money

Beginners learn from tips and luck, practise with real savings, and rarely get honest feedback on why a decision was good or bad. Testing a trading idea usually needs programming skills most learners don't have.

Our solution

Practice on real history, graded on decision quality

Candle-by-candle replay of real Nepal Stock Exchange data with virtual money.
Scores for patience, position sizing and risk control, not just profit.
Plain-English ideas turned into tested rules by a private, on-premise AI.
A structured 8-week program with lessons, quizzes, missions and leaderboards.

Feature explorer

Everything we built

Practice on real market history

A replay of real NEPSE price history, one candle at a time, with virtual money.

Candle-by-candle replayStep through or play at 1×, 2× or 4× speed. The future stays hidden.

Realistic ordersBuy, sell and close with share sizing, stop and target, and the real round-trip cost shown before every trade.

Live position panelEntry, current value, unrealised P&L and drawdown, updated every candle.

Reflection built inThoughts, hypotheses and key insights are recorded in each session.

Trading cockpit: a live replay with an open position

Practice with a purpose

Every session can carry a challenge that trains a specific skill.

246 NEPSE stocksReal daily candles, with adjustable session length.

Skill challengesFree play, max 3 trades, low drawdown (keep losses under 5%) and prediction challenges.

Missions & rewardsA first mission and XP rewards guide new learners into their first session.

Anonymous by defaultLearners get a fun anonymous trader name, so no email is needed to start.

New session: stock, length and challenge

Describe an idea in plain English

Type a belief about the market and AI turns it into a precise rule that can be tested.

Natural language input"Buy after the price drops 2% over the last 20 candles, then sell 20 candles later."

Private, on-premise AIA self-hosted open model, so learner data never goes to a third-party AI service.

The AI drafts and the code decidesEvery model output is checked by deterministic code before anything runs.

Guided examplesOne-click starting ideas such as RSI below 30 or a 20-candle high.

Results that tell the truth

Backtests on real data, with the costs and caveats spelled out.

Real NEPSE costsBroker commission, SEBON fee and DP charge are applied on both buy and sell.

Compared against luck"Beat 9 of 10 random traders", with an honest "too few trades to judge" warning when it applies.

Idea → Result → EvidenceA clear three-step flow from belief to proof.

Every trade inspectableTrade markers on the chart, plus win rate, drawdown and per-trade economics.

A structured 8-week program

From "Your First Experiment" to "Does it still work later, and elsewhere?".

Weekly unlocksEach week builds one skill and unlocks the next.

Progress trackingWeeks submitted, sessions played and a full record of a learner's thinking.

GamificationXP, levels, streaks, achievements and a fair leaderboard that one lucky run can't top.

Versioned editionsCohorts can be pinned to a specific edition of the course.

Lessons for complete beginners

A 7-day warm-up plus a searchable lesson library.

NEPSE FoundationsA 7-day warm-up: what the stock market is, what NEPSE is, reading your first stock.

Lesson library22 short lessons on candlesticks, order types, FOMO, revenge trading, trends and more.

Quizzes & missionsCheck understanding and apply it straight away.

Works like an appInstallable on phones, offline-aware and optimised for iPhone.

Verified, not just claimed

A public validation page explains exactly what is verified, and what isn't yet.

Deterministic engineThe same inputs always give a byte-identical replay, checked by stress tests on every core change.

5,300+ automated testsUnit, property-based, UI and end-to-end browser tests run in CI.

Secure by defaultPasswordless and Google sign-in, admin MFA, rate limiting, strict security headers and user data export or deletion.

Tutor toolsStudent roster, learning-insight panels, messaging and an audit log for cohort teaching.

On mobile

The full academy, in your pocket

Trade: set up a practice session — Practice setup

Research Desk: idea to rule — Research Desk

Architecture

How it fits together

Learner

Web & mobile PWAReact 19

Push & email

Edge

Cloudflaresecure public access

nginxstrict security headers

Application

Replay enginedeterministic

FastAPI160 endpoints

Scoring & gamification

Intelligence

Private LLMon-premise

Rule validatorcode-enforced

Backtest engine

Data

PostgreSQL 16

Market data serviceour NepseScanner platform

Daily backups

Technology

Tech stack

Backend

Python 3.12FastAPIPydantic v2SQLAlchemy 2AlembicPostgreSQL 16

Frontend

React 19TypeScriptViteRechartsCustom candle chartsPWA

AI

Ollama (self-hosted)Qwen 2.5NL → structured rulesOutput validation

Infrastructure

DockernginxCloudflaresystemd timersGitHub ActionsSentry

Quality

pytestHypothesisVitestPlaywrightmypyESLint

Integrations

Google sign-inResend emailWeb PusheSewa

Engineering quality

Built to be trusted

Deterministic core

The same inputs give identical results every time, checked by stress tests on every core change.

5,300+ automated tests

Unit, property-based, UI and end-to-end browser tests run in CI.

Survives reboots

Services and scheduled jobs recover automatically, and missed backups catch up.

Security first

MFA, strict content-security policy, rate limits and regular vulnerability remediation.

Documented & governed

Build spec, disaster-recovery plan, release checklist and dated audits.

Safe releases

Deploy script with rollback tags and uptime checks.

What this means for your business

The same capabilities, applied to your problems

Private AI on your serversLLM features that keep sensitive data in-house, with code-enforced guardrails.

Training & e-learning platformsCourses, assessments, gamification and tutor analytics for your staff or customers.

Plain-language interfacesLet your team describe what they want in plain words and have AI turn it into structured actions.

Next project

NepseScanner →

View