Real-World AI Engineering
For High School Students.

Don’t settle for “toy” projects. Join the HighScores.ai core team for 8 weeks to build, deploy, and scale production features using the industry’s most advanced AI tech stack.

PROGRAM DIRECTOR

Message from the Founder

Guided by an EdTech Architect.

“This isn’t a coding camp. It’s a technical fellowship where high-achieving students work on the same architectural bottlenecks I solve every day at HighScores.ai. My goal is to help you build a portfolio that admissions officers at Stanford and MIT can’t ignore.”

DIRECTOR OF ENGINEERING

Founder of HighScores.ai | CS Educator | AP Computer Science Specialist

The Project Explorer

Select a core track. Each project represents a high-stakes engineering bottleneck at HighScores.ai.

OPTIMIZATION

DATA SCIENCE

01. The AI Academic Architect

Static study plans are dead. You will engineer a Constraint-Satisfaction Engine that takes student diagnostic data and outputs an 8-week optimized learning trajectory based on the student’s “Point-per-Hour” return.

  • The Technical Challenge

    Balancing SAT unit weightage against student fatigue and daily time constraints using weighted scoring models.

  • Resume Impact

    "Architected an automated pedagogical engine utilizing recursive logic and optimization algorithms."

# Optimization Logic Mockup
plan = calculate_path(score_gap, hours_available, unit_weights)
optimized_schedule = priority_queue.fit(plan)
Pandas

LIBRARY

JSON

STRUCTURE

LLM Gen

SYNTHESIS

GENERATIVE AI

LATEX SUPPORT

02. Synthetic Content Engine

Scale our question bank by 10x. You will build a Prompt-Chaining Pipeline that “clones” existing math problems, changing the narrative context and variables while ensuring the pedagogical difficulty remains identical.

  • The Technical Challenge

    Preventing AI hallucinations in mathematical proofs and outputting valid LaTeX formatted code consistently.

  • Resume Impact

    "Engineered a Generative AI pipeline utilizing few-shot prompting and symbolic verification agents."

INPUT: LINEAR EQUATION QUIZ

“John has 5 apples and buys 3 more…”

“The rocket burns 5L of fuel and adds 3L…”

NLP

COGNITIVE AI

03. The Cognitive Error Lab

Why did the student miss the question? You will build a Diagnostic Classifier that analyzes “distractors” (wrong answers) to identify the mental misconception (e.g., calculation error vs. conceptual misunderstanding).

  • The Technical Challenge

    Classifying qualitative error patterns using LLMs and mapping them to a taxonomy of cognitive biases.

  • Resume Impact

    "Developed an NLP-driven error profiler to automate qualitative student feedback at scale."

Wrong

“Calculation Error”