Call For Tasks
The IOAI International Scientific Committee (ISC) invites AI researchers worldwide to contribute to the design of competition tasks for IOAI 2027.
The IOAI is the premier annual AI competition for high school students, focusing on topics including but not limited to machine learning, computer vision, and natural language processing
The competition consists of three stages:
- At-Home Round: Students work on three tasks over one month at home. This round is educational and does not affect the final results.
- Contest 1: Continuation of the at-home tasks.
- Contest 2: Features novel tasks, unrelated to the at-home stage.
Contest 1 and Contest 2 are individual contests, which will take place during IOAI 2027 in Singapore (July 4ā10, 2027).


Task requirements
The proposed tasks must:
- Be aligned with the IOAI Syllabus.
- Be original, unseen by IOAI 2027 contestants, and not reused from other competitions.
- Be tailored to the capabilities of top high school students representing national teams, striking a balance between theoretical understanding and practical application.
You can find IOAI 2024ā2026 tasks here.
This year we invite contributions for: fully developed IOAI-style tasksĀ and/or theoretical tasks.
Non-theoretical submissions should follow the task template and include a full task statement and code for data loading, baseline solution and evaluation function. Authors of complete tasks selected by the ISC for use in IOAI 2027 will receive recognition at IOAI 2027 and will be invited to IOAI 2027 as Jury members.
We also encourage authors to follow the comprehensive task-development guide here.
This year, we welcome more theoretical tasks aimed at a deep understanding of the areas listed in the Syllabus. These could be regular pen-and-paper tasks, machine learning problems that require first-principles understanding of ML fundamentals, or a mix of both.
We will consider proposals from the community and may incorporate parts of them into the regular IOAI-style tasks.
Submission process
Complete tasks and task ideas should be submitted through this Google form.
Task selection criteria
Task submissions will be evaluated based on the following criteria:
Hammer resistance
Good tasks are not easily solved with off-the-shelf, general-purpose methods (hammers) or brute force, but instead require problem-specific adaptations and creative thinking.
Multiple insights and partial solutions
Good tasks require multiple insights for a complete solution, and have scope for progressively rewarding partial solutions.
Educational value
Good tasks reward rigorous understanding, teach useful skills, and inspire students to learn more.
Fast iteration cycles
Good tasks do not require lengthy training or processing steps, so more time can be spent on problem-solving instead. Tasks should execute in under 10 minutes on a T4 GPU card and solution files should not exceed 1 GB.
Suitable data
Good tasks rely on data of adequate size, quality and alignment with the task (unless the task is explicitly about tackling issues with data properties).
Originality
Good tasks are inventive, focus on atypical applications, go beyond standard problem settings.
For any questions or clarifications regarding the Call for Tasks, please contact the ISC at isc@ioai-official.org
