Intelligent Agent for a Smart-Thermostat Pricing-Aware Schedule
Overview
What this challenge is about.
Design an intelligent agent with: perception (read sensor history), basic learning (cluster comfort intervals from 7 days of observations), decision-making (schedule heating to minimize cost given tariff windows + predicted comfort intervals), action (output the next 24h thermostat schedule). Implement in Python with simulated household data (provided: 30 days of sensor + tariff data for 12 households). Evaluate: % bill reduction, comfort-violation count (degrees-hour deviation), schedule-stability (no whiplash). Compare against a fixed-schedule baseline. Deliver: agent code + tests, evaluation report (6 pages), 12-household comparison dashboard (Streamlit), 1-page memo on which agent decisions a household should be able to override.
The Brief
What you'll do, and what you'll demonstrate.
Build an intelligent agent that cuts thermostat bills by 12 percent without comfort violations across 12 simulated households.
Earning criteria — what you'll demonstrate
- Design a complete intelligent agent (perception, learning, decision, action)
- Apply basic unsupervised learning (clustering) to observed behavior
- Optimize decisions under multiple-objective constraints (cost vs. comfort)
- Evaluate agent performance with realistic simulated household data
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Product Manager
PMs on smart-home products need this agent-design fluency to scope adaptive features that don't quietly burn user trust on a single bad comfort violation.
This challenge sharpens
- intelligent-agents
- algorithm-evaluation
- basic-learning