Build a Pricing-Aware Scheduling Agent for Smart Thermostats
Overview
What this challenge is about.
Build a scheduling agent that reads sensor data, learns comfort patterns, and minimizes heating cost across tariff windows. Earn your verifiable certificate.
The scenario
The company is a direct-to-consumer smart-home brand that sells internet-connected thermostats to South Korean households, where utilities charge different electricity rates depending on the hour of the day.
The Brief
What you'll do, and what you'll demonstrate.
Households overpay because their thermostats heat during expensive peak hours and ignore each home's real comfort routine, so the thermostat must learn comfort windows and shift heating to cheaper hours without ever leaving a home uncomfortable.
Earning criteria — what you'll demonstrate
- Decompose an intelligent agent into perception, learning, decision, and action layers that pass clean interfaces between each other
- Apply unsupervised clustering to infer comfort intervals from noisy sensor history
- Formulate and solve a cost-minimization schedule under hard comfort and price-window constraints
- Evaluate an agent honestly against a baseline using multiple, sometimes-competing metrics
- Translate algorithmic decisions into human-override surfaces a real product team can ship
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Aligned coursework coming soon.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
- Intelligent Agents
Apply intelligent agents to solve real industry problems and demonstrate production-level capability.
- Basic Learning
Apply basic learning to solve real industry problems and demonstrate production-level capability.
- Python Programming
Apply python programming to solve real industry problems and demonstrate production-level capability.
- Algorithm Evaluation
Apply algorithm evaluation to solve real industry problems and demonstrate production-level capability.
- Optimization
Apply optimization to solve real industry problems and demonstrate production-level capability.
- Knowledge Representation
Apply knowledge representation to solve real industry problems and demonstrate production-level capability.
Careers
Career paths this challenge builds toward
Completing this challenge demonstrates skills that transfer directly to these roles:
Machine Learning Engineer
Designing the perception-learning-decision-action loop and clustering comfort intervals from raw sensor history mirrors how ML engineers turn messy device data into shipped, measurable model-driven behavior in production systems.
This challenge sharpens
- intelligent-agents
- basic-learning
- python-programming
Backend Software Engineer
Extending a modular agent skeleton with tested interfaces and a deterministic simulation harness is the same discipline backend engineers use to build reliable, testable services that other teams depend on.
This challenge sharpens
- python-programming
- algorithm-evaluation
- knowledge-representation
Optimization Engineer
Formulating heating as a cost-minimization problem under hard comfort and price-window constraints, then proving the gains against a baseline, is exactly the modeling-and-evaluation work optimization engineers do in energy and logistics.
This challenge sharpens
- optimization
- algorithm-evaluation
- intelligent-agents