AI Energy Optimization for Data Centers
Publicada el 2026-07-24
Descripción de la oferta
I’m launching an AI-based product aimed at reducing the electricity bill of large-scale data centers by making smarter, data-driven decisions about when and how equipment draws power. My focus is strictly on cutting energy consumption, not on broader resource allocation or cooling alone. The core of the solution will rely on machine learning algorithms that learn from live telemetry (power meters, workload logs, environmental sensors) and then recommend or automatically trigger actions such as server throttling, dynamic workload shifting, or turning on low-power modes. I already have access to historical datasets and can arrange remote access to a small test lab for validation, but I need an experienced partner who can take the concept from raw data to a production-ready model and lightweight dashboard. What I’m looking for • End-to-end ML workflow: data cleaning, feature engineering, model selection, training, and continuous learning pipelines • Deployment strategy for real-time inference, ideally containerised so it can sit inside existing on-prem infrastructure • Clear metrics demonstrating kWh savings and model accuracy, validated against a baseline Please submit a detailed project proposal outlining the approach, timeline, and any prior experience with energy optimisation or data-center telemetry. I’ll review proposals on how convincingly they translate energy-saving theory into measurable results and how well risks around data quality, latency, and model drift are addressed. I’m ready to start as soon as the right plan is on the table, and I’m open to iterative milestones so we can showcase early wins before full rollout.
Skills
Fuente original: freelancer