Hybrid AI Research Engineer (Post-doc) - Digital Twin & RL

Hewlett Packard Enterprise; Galway; Hybrid; EUR 65,000 - 90,000; Full time

Job summary

Hewlett Packard Labs in Galway is seeking a Machine Learning/AI Research Engineer (Post-doc) to develop AI-driven control for data-centre digital twins and live-test environments. The role spans ML, software engineering and experimentation, with collaboration across international partners.

You will build digital-twin software, forecast IT load and cooling signals, and train multi-objective RL policies while validating through hardware-in-the-loop and controlled trials.

Qualifications

• PhD degree, or Master's degree with equivalent research and innovation experience.
• Proven record of innovation in deep learning, scientific computing, engineering computation, or machine-learning applications.
• Evidence of independent research leadership, for example via publications in leading conferences or journals.

Responsibilities

• Build data and workload pipeline, telemetry ingestion, time-series data management and baseline datasets.
• Develop modular digital-twin software combining telemetry, models, cooling and storage models, and constraints.
• Create probabilistic forecasts for IT load, cooling demand and water/heat signals.
• Research and evaluate multi-objective reinforcement-learning policies balancing cost, carbon, water, and durability within safety limits.
• Implement a price-response interface with auditable flexibility commitments and uncertainty.

Skills

Python

PyTorch

Reinforcement learning

Digital twins

Time-series

System-level modelling

Education

PhD or MSc with equivalent research experience

Job description

Hewlett Packard Labs in Galway is seeking a Machine Learning/AI Research Engineer (Post-doc) to develop AI-driven control for data-centre digital twins and live-test environments. The role spans ML, software engineering and experimentation, with collaboration across international partners.

You will build digital-twin software, forecast IT load and cooling signals, and train multi-objective RL policies while validating through hardware-in-the-loop and controlled trials.