← Modelplay
The challenges

Pick a track.
Commit to the reps.

Every challenge runs entirely in your browser — the brief on the left, a real editor and live output on the right, on a focus timer. Real datasets, real scikit-learn, real AI models. Finish a track and walk away with a portfolio that proves it.

71hands-on challenges
4guided tracks
3skill levels
$0in-browser, no setup
Category
Status
27 tracks
Multi-day tracksOne focused build a day
CertificationLive

Noob Dev Engineer

A sealed, timed exam: 20 intermediate-to-advanced questions across ML and AI, 10 minutes each. Pass to earn the Noob Dev Engineer badge.

20questions
10 mineach, sealed
ML + AIcombined
Take the exam →
Practice by level10 challenges each · 5 ML + 5 AI
Career tracks30-day role journeys · coming soon
Soon

30 Days of Data Analytics

Become a Data Analyst

From spreadsheets to SQL to dashboards — 30 days of turning raw data into decisions a business can act on.

Spreadsheets to SQLClean & shape messy dataExploratory analysis+3 more
30 daysPreview →
Soon

30 Days of Data Science

AI & Data Science track

The full pipeline — wrangling, modelling, evaluation, and communicating results — across 30 hands-on days.

Pandas & data wranglingStatistics that matterFeature engineering+3 more
30 daysPreview →
Soon

30 Days of Data Engineering

Become a Data Engineer

Build the pipelines everything else depends on — ingestion, warehousing, and orchestration over 30 days.

Data modelling & schemasBatch & streaming ingestionWarehouses & lakes+3 more
30 daysPreview →
Soon

30 Days of SQL

Query like a pro

From SELECT to window functions and query tuning — the language every data role runs on, one day at a time.

SELECT, filter & sortJoins across tablesAggregation & grouping+3 more
30 daysPreview →
Soon

30 Days of Python for Data

Python, the data way

NumPy, pandas, and clean, reusable code — the Python foundation every analyst and scientist needs.

Python essentialsThink in NumPy arraysWrangle with pandas+3 more
30 daysPreview →
Soon

30 Days of Statistics

The math behind the models

Probability, inference, and hypothesis testing — the intuition that keeps your conclusions honest.

Distributions & probabilitySampling & the CLTConfidence intervals+3 more
30 daysPreview →
Soon

30 Days of Deep Learning

Neural nets from scratch

Build up from a single neuron to full networks — forward passes, backprop, and training that converges.

Neurons & activationsForward & backpropTraining & optimization+3 more
30 daysPreview →
Soon

30 Days of NLP

Teach machines language

From tokens and TF-IDF to transformers — the techniques that turn raw text into understanding.

Tokenization & cleaningBag-of-words & TF-IDFWord embeddings+3 more
30 daysPreview →
Soon

30 Days of Computer Vision

Make machines see

Images as data — filters, features, and convolutional networks for classification and detection.

Images as arraysFilters & featuresImage classification+3 more
30 daysPreview →
Soon

30 Days of Generative AI

Build with foundation models

Text, images, and beyond — how generative models work and how to build real products on them.

How generation worksWorking with LLM APIsImage generation+3 more
30 daysPreview →
Soon

30 Days of LLM Engineering

Production-grade LLM apps

Beyond the prompt — RAG, evaluation, caching, and cost control for LLM systems that hold up in production.

Prompting patternsRetrieval-augmented generationStructured & tool calling+3 more
30 daysPreview →
Soon

30 Days of AI Agents

Tools, memory & autonomy

Give models hands — planning, tool use, memory, and multi-step agents that actually get work done.

Anatomy of an agentTool use & function callsPlanning & reasoning+3 more
30 daysPreview →
Soon

30 Days of Prompt Engineering

Get exactly what you want

A daily drill in the patterns that reliably steer LLMs — few-shot, chain-of-thought, structure, and evals.

Anatomy of a promptFew-shot examplesChain-of-thought+3 more
30 daysPreview →
Soon

30 Days of MLOps in Production

From notebook to live service

The full lifecycle — packaging, CI/CD, monitoring, and retraining — for models that survive contact with prod.

Package & version modelsServe behind an APICI/CD for ML+3 more
30 daysPreview →
Soon

30 Days of Time Series

Forecast the future

Trends, seasonality, and forecasting — from classic models to ML approaches for data that moves through time.

Trends & seasonalityStationarity & differencingARIMA family+3 more
30 daysPreview →
Soon

30 Days of Recommender Systems

Personalize everything

The engines behind “you might also like” — collaborative filtering, embeddings, and ranking, day by day.

Recommendation basicsCollaborative filteringMatrix factorization+3 more
30 daysPreview →
Soon

30 Days of Reinforcement Learning

Learn by trial & reward

Agents that learn from consequences — rewards, policies, and Q-learning, built up over 30 days.

Agents, states & rewardsMarkov decision processesQ-learning+3 more
30 daysPreview →
Soon

30 Days of Big Data

Scale past one machine

When data won’t fit in memory — distributed processing with Spark and the patterns of big-data pipelines.

Why distributed computingMapReduce thinkingSpark fundamentals+3 more
30 daysPreview →
Soon

30 Days of Business Intelligence

Dashboards that drive decisions

Metrics, modelling, and visualization — turn warehouse data into dashboards leaders actually use.

Defining good metricsDimensional modellingBuilding dashboards+3 more
30 daysPreview →
Soon

30 Days of AI & ML Security

Break it before they do

Adversarial attacks, prompt injection, and model hardening — securing AI systems end to end.

The AI threat modelAdversarial examplesPrompt injection & jailbreaks+3 more
30 daysPreview →

Want a lab that isn’t here yet?

Tell us what you want to learn next, or just say hi. We ship new labs and tracks all the time — your request helps decide what’s next.