DataCenter_EnergyOptimization
The solution will be evaluated on its ability to accurately identify and propose optimizations for simulated data center scenarios (e.g., energy spikes, potential hardware failures), its effective use of Gemini 2.5 Pro's capabilities, the quality of agent collaboration, and the demonstrable efficiency gains (simulated).
Evaluation overview
How the linked challenge is judged: tasks, benchmarks, and criteria count.
Task templates
Inputs and expected outputs.
DataCenter_EnergyOptimization
Evaluates the agent system's ability to analyze simulated power consumption data and propose specific actions to reduce energy usage without impacting performance.
{ "telemetry": { "current_power_draw_mw": float, "server_rack_temps_c": {"rack_A": float, "rack_B": float}, "utilization_percent": {"cpu": float, "gpu": float}, "workload_type": "string" }, "energy_price_per_kwh": float, "optimization_goals": ["string"] }
{ "proposed_actions": ["string"], "estimated_savings_kwh": float, "reasoning_trace": ["string"], "agent_collaboration_summary": "string" }
DataCenter_PredictiveMaintenance
Assesses the system's ability to identify potential hardware failures from multimodal sensor data and historical trends, then recommend proactive maintenance.
{ "telemetry_history": [{"timestamp": "datetime", "fan_rpm": int, "vibration_level": float, "power_draw": float}], "current_multimodal_input": {"image_url": "url", "audio_data": "base64_string", "logs": ["string"]}, "maintenance_thresholds": {"vibration_high": float, "fan_low": int} }
{ "potential_failure": "string|null", "confidence_score": float, "recommended_action": "string", "predicted_failure_date": "date|null", "agent_decisions": ["string"] }