Classes 11-12 · Year two • Cho Oyu SummitEngineering-grade

Decide what system should exist at all.

Cho Oyu is not 'Pumori with more AI.' The learner now investigates an ambiguous problem, identifies stakeholders and constraints, decides what should and should not be automated, chooses the technical approach, engineers the system, evaluates it and defends the consequences of those decisions.

“I can independently investigate, build, evaluate and defend complex computational work.”

Who owns whatThe learner increasingly owns the problem itself: investigation, stakeholders, requirements, approach, architecture, implementation, evaluation, human judgment and defense.

“I can independently engineer a serious system for a defined problem.”

“I can determine what problem is worth solving, what system should exist, build it under uncertainty and defend the complete decision chain.”

What changes this year

Problem discovery and requirements become learner-owned.
The learner distinguishes symptoms from underlying problems and identifies stakeholders with competing needs.
Success criteria, unacceptable failure and non-functional requirements are explicit.
The learner may correctly conclude that a problem should not be solved primarily with AI.
Architecture is designed under ambiguous requirements and trade-offs rather than supplied.
Data/system design includes storage, pipelines, models, generative AI, agents and deterministic components only where justified.
Evaluation includes deterministic correctness, AI/model evidence and whether the product actually helps the intended user.
Independent technical learning becomes a first-class capability: unfamiliar libraries, APIs, formats and codebases can be investigated and validated.
The final defense covers technical, product, human and evidence decisions.

The destination is technical judgment

Try the thinking

A minute of Cho Oyu Summit thinking.

An illustrative interaction, not a lesson from the curriculum — but the reasoning is the reasoning.

The brief says 'build AI.' The first job is to question the brief.

Interactive example

Cho Oyu begins where a normal coding assignment ends: with ambiguity.

The brief, as received

“Our school buses leave late. Build an AI system to decide when each bus should depart.”

Before choosing a technology, investigate. Pick up whatever you think matters — nothing here is scored.

Stakeholders

Possible causes

Constraints

Nothing gathered yet.

Which is the strongest first problem statement?

What they produce

Problem investigation / stakeholder evidence
Requirements and success/failure criteria
Alternative solution strategies
Learner-authored architecture
Algorithm/data-structure rationale
Data/model/AI evaluation strategy
Reliability/security/privacy analysis
Human-control and automation decisions
Technical documentation
Full technical/product defense

What you should be able to see

Evidence that the problem was investigated rather than merely accepted
Requirements, constraints and unacceptable failure
Several plausible solution strategies compared
Why AI/ML/agents were—or were not—appropriate
Architecture and data-system decisions
Testing/evaluation designed by the learner
A technical and human-risk analysis
A full defense of the decision chain

Support matters. A heavily guided success is not reported as independent mastery.

Twelve Projects. One climb.

A Band is not a pile of unrelated activities. Each standard Band contains 12 Projects arranged so that capability is established, combined, transferred and increasingly owned.

P1–P4Establish / Acquire
P5–P8Combine / Extend
P9–P12Transfer / Own

Illustrative synthesis example

Investigate → formulate → compare approaches → architect → engineer → evaluate → defend

The student receives no clean technical brief. They establish that the problem exists, define stakeholders and success, compare plausible approaches, design a substantial architecture, build only the complexity the problem earns, and defend both what they automated and what they deliberately kept human.

Design archetype: Independent Intelligent Systems Capstone

Illustrative capstone architecture from current Band design; final Project title/domain must come from approved curriculum.

Where Cho Oyu Summit lands on each capability

These are Band end-states, not lesson titles. Follow any of them across all eight years in the Capability Telescope.

Algorithms & problem solving: Select or adapt algorithm/data-structure families under real constraints and defend performance, maintainability and system trade-offs.
Programming & representation: Independently map unfamiliar codebases, follow data/control flow and integrate unfamiliar technical components.
Debugging: Diagnose failures involving architecture, data, configuration, dependencies, asynchronous behavior and interacting components under ambiguity.
Testing & reliability: Design the full deterministic + AI-system evaluation strategy, including unacceptable failure and product validation.
Data & evidence: Design and defend the data system: sourcing, joins, storage, sampling, uncertainty, provenance and what the evidence can legitimately support.
AI & machine learning: Choose whether AI belongs in the solution at all; architect and evaluate ML/generative/agentic components under stakeholder, cost, privacy, reliability and human-control constraints.
Systems & architecture: Architect interacting subsystems under ambiguous requirements, competing constraints and failure propagation—and defend why this architecture fits the problem.
Technical judgment: May correctly conclude that AI is unnecessary or that a proposed automation should not exist; defend the full technical and human decision chain.
Security, privacy & human-centered computing: Own stakeholder analysis, appropriate automation, human control, privacy/security and sociotechnical consequences.
Explanation, transfer & technical defense: Defend the problem framing, requirements, technical paradigm, architecture, evidence, risks, human-control choices and unresolved limitations.

Open the Capability Telescope

The summit is not 'I can use AI.'

It is: I can investigate a real problem, decide what computational system ought to exist, engineer it, evaluate it honestly and defend the consequences of the choices I made.

What prepared them: Pumori Summit See the whole climb

The whole climb