USA–North America AI Olympiad · 2027 Cycle
USAAIO Preparation Course
A from-scratch curriculum taking a motivated middle-schooler from zero Python to Olympiad-level machine learning and deep learning. 从零开始，带学生从 Python 入门到奥赛级别的机器学习与深度学习。
Goal: Qualify for Round 2
10+ hrs / week
Python from scratch
~30
weeks until Round 1 (late Jan 2027)
12
Harper's age on first attempt
Start here
The competition in one glance
Round 1
Open to everyone. Late Jan 2027, 3 hours, online & proctored. Google Colab, CPU only. Foundations: math, Python, classical ML, intro deep learning.
Round 2
Qualify via Round 1. Mar–Apr 2027, at MIT. Covers everything on the syllabus; GPUs allowed (Colab L4).
Camp → Team USA
June 2027 at Harvard. Top campers selected for IOAI / IAIO international teams.
Selection funnel: 2025 vs 2026
Pass rates at each stage. 各阶段通过率对比。 2026 Round 1 total and camp/team counts not yet officially published — marked as estimates.
| Stage / 阶段 |
2025 |
2026 |
Pass rate / 通过率 |
| Round 1 participants / 第一轮人数 |
393 |
not published |
— |
| → Round 2 / 晋级第二轮 |
76 (cutoff 53/300) |
~200 |
19.3% (2025) · n/a (2026) |
| → Camp / 集训营 (Round 3) |
12 |
~12 (est.) |
15.8% of R2 · ~6% of R2 (2026) |
| → Team USA (IOAI) / 美国队 |
8 |
8 (est.) |
67% of camp |
| Overall: R1 → Team / 总录取率 |
8 / 393 |
— |
~2.0% (2025) |
Sources: usaaio.org · ioai-official.org (2026 Round 2 at Harvard & MIT, ~200 advanced) · competitionsciences.org
How this course is built 🌙
Materials are developed automatically overnight (≈12am–5am) when the laptop is free. Each morning, new lecture notes and practice sets appear, and Harper's progress page gets an updated summary. 课程材料在夜间自动开发，每天早上会有新内容。