Build Dynamic BMS Telemetry and Office HVAC Optimization Agent using Vercel AI SDK
Improving net operating income (NOI) in distressed commercial office assets requires reducing building operating expenses. Developers will build an automated Building Management System (BMS) energy optimization agent using the Vercel AI SDK. The agent ingests real-time HVAC occupancy telemetry, local energy tariffs, and weather forecasts. It dynamically adjusts building setpoints and operational schedules to minimize peak energy charges, lowering overall operating costs.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an automated building operations agent using Vercel AI SDK to optimize HVAC telemetry and lower office operational expenses.
How work is evaluated
Evaluates calculated energy cost savings and verification of indoor comfort threshold preservation.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Comfort Boundary Validation
Ensures setpoint recommendation is within comfortable limits (<=76 F)
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Savings Optimization Precision
Accuracy of estimated percentage savings • target: 0.92 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Process time-series BMS IoT telemetry streams using Vercel AI SDK
Build automated tool callbacks for adjusting HVAC temperature setpoints and airflow rates
Optimize energy usage during peak tariff windows based on forecasted occupancy
Calculate monthly operating cost reductions for real estate asset dashboards
Reference links and supporting material
Dataset of 100 hourly building BMS telemetry records including indoor/outdoor temp, power draw, and occupancy.
Challenge setup
Evaluates calculated energy cost savings and verification of indoor comfort threshold preservation.
Challenge input
JSON with occupancy_pct, outdoor_temp_f, peak_tariff_active, current_setpoint_f
AI SDK
Required framework for TypeScript agents.
Evaluated output
JSON with recommended_setpoint_f, estimated_cost_savings_pct, comfort_maintained (boolean)
- Ensures setpoint recommendation is within comfortable limits (<=76 F)
- Accuracy of estimated percentage savings • target: 0.92 • range: 0-1
- Benchmark: BMSOptimizationBench
- Savings Optimization Precision target: 0.92
- 1 public reference case
- JavaScript execution harness
- javascript sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Vercel AI SDK · Required
- Zep · Optional
- Fiddler AI · Optional
- Vercel AI SDK · Required
Evaluation contract
- Comfort Boundary Validation · Weight 1
- Savings Optimization Precision · Weight 1
Metrics
- Savings Optimization Precision · Target 0.92
Benchmarks
- BMSOptimizationBench
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with occupancy_pct, outdoor_temp_f, peak_tariff_active, current_setpoint_f
Output: JSON with recommended_setpoint_f, estimated_cost_savings_pct, comfort_maintained (boolean)
Execution harness: JavaScript
Evaluation tools
- AI SDK · Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.