STEMist hackathon winners and other participants, claim your gifts here.

USACO

Home ยป USACO
Contact Master Ji

AI

TinyGPT Builder Camp

Build Your Own Mini Language Model

Best for
Students interested in AI, ChatGPT-style models, and language models

Duration
5 days

Class length
2 hours/day

Price
$129 Early Bird / $159 Regular

Date
20, 21, 22, 23, 24 (9 am pst-11:30 am pst)

What students will learn

  • What a language model is
  • How language models predict the next token
  • What a text dataset looks like
  • Characters, words, tokens, and vocabulary
  • How text is converted into numbers
  • Tokenisation and encoding
  • Input and target creation for training
  • Basic idea of embeddings
  • Basic idea of model architecture
  • What training means
  • What loss means in language modelling
  • How a model improves during training
  • How to generate text using prompts
  • Temperature and randomness in generated text
  • Why small models make mistakes

Hands-on activities

  • Explore a text dataset
  • Preprocess text for training
  • Convert text into tokens
  • Assemble a small educational language model using guided code
  • Train the model in Google Colab
  • Track loss during training
  • Save and run the trained model
  • Generate text using custom prompts
  • Experiment with temperature and output length
  • Final mini showcase

Outcome

Students will build and run a small educational language model step by step. They will understand the pipeline behind language models: dataset to tokenisation to model to training to inference.

Requirements

  • Basic Python knowledge required
  • Machine Learning Explorer Camp recommended
  • Laptop or desktop
  • Google account
  • Google Colab access
  • Stable internet connection
  • No powerful laptop or GPU required

Note

This is a guided AI builder camp. Students do not need to understand every mathematical detail. They will work with prepared code, simple explanations, and hands-on experiments.

Contest

USACO Bronze Sprint Camp

Competitive Programming and Problem Solving

Best for
Students preparing for USACO Bronze or coding contests

Duration
5 days

Class length
3 hours/day

Price
$99 Early Bird / $129 Regular

Date
27, 28, 29, 30, 31 July (9 am pst-11:30 am pst)

What students will learn

  • How to read and understand contest problems
  • Input/output format used in coding contests
  • Simulation-based problem solving
  • Arrays and lists
  • Strings and character operations
  • Frequency counting
  • Brute force techniques
  • Sorting-based problem solving
  • Greedy thinking
  • Prefix sums
  • Basic time complexity: O(n), O(n log n), O(n^2)
  • How to debug contest code

Hands-on activities

  • Around 5-10 Codeforces/LeetCode-style problems
  • Around 5 USACO Bronze-level problems
  • Guided problem solving
  • Independent practice problems
  • Discussion of different approaches

Outcome

Students will learn the core techniques needed for USACO Bronze and become more confident in solving competitive programming problems.

Requirements

  • Basic Python knowledge
  • Comfort with loops, lists, strings, and conditions
  • Laptop or desktop
  • Internet connection

ML

Machine Learning Explorer Camp

Visualising Machine Learning with Python

Best for
Students with basic Python knowledge

Duration
4 days

Class length
2 hours/day

Price
$49 Early Bird / $69 Regular

Date
15, 16, 17, 18 July (9:00 am pst -11:30 am pst)

What students will learn

  • What machine learning is and where it is used
  • How machines learn from data
  • Datasets, rows, columns, features, and labels
  • Basics of NumPy, Pandas, and Matplotlib
  • Visualising data using simple charts
  • Regression: predicting numbers
  • Classification: predicting categories
  • Train/test split
  • Model training and prediction
  • What loss and loss function mean
  • Accuracy, error, and model performance
  • Basic idea of overfitting
  • How text can be converted into numbers for AI models

Hands-on activities

  • Explore real datasets
  • Visualise data using graphs
  • Build a simple house price prediction model
  • Build a simple classification model
  • Compare actual vs predicted results
  • Understand how training improves a model

Outcome

Students will understand the basic workflow of machine learning: data to training to prediction to evaluation. This camp also prepares students for the TinyGPT Builder Camp.

Requirements

  • Basic Python knowledge
  • Laptop or desktop
  • Internet connection
  • Google Colab access recommended

Free

Python Spark Camp

Introduction to Programming with Python

Best for
Beginners

Duration
2 days

Class length
2 hours/day

Price
Free

Date
13 and 14th July (9:00 am pst -11:30 am pst)

What students will learn

  • What programming is and how Python works
  • Variables, numbers, strings, and basic data types
  • Taking input from the user
  • if, elif, else conditions
  • Loops using for and while
  • Lists and basic list operations
  • String operations and simple logic building
  • Writing small beginner-friendly Python programs

Hands-on activities

  • Number checker
  • Marks/grade calculator
  • Simple quiz game
  • Guess the number game
  • Basic list and string problems

Outcome

By the end of this camp, students will be able to write simple Python programs and understand the basic building blocks of coding.

Requirements

  • Laptop or desktop
  • Internet connection
  • No prior coding experience required