Head of AI Infrastructure (MLOps)

Head of MLOps & AI Infrastructure

This role bridges research and production, owning the full model lifecycle — from experimentation to operational deployment.

The position includes building scalable MLOps infrastructure, automating training and inference workflows, and continuously improving performance and reliability in distributed systems.

Key Responsibilities

• Design and lead end-to-end MLOps processes:

Data → Training → Evaluation → Deployment → Monitoring

• Develop model training pipelines, including experiment management and comparative evaluations

• Implement and manage Model Registry, versioning, and artifact management (MLflow / DVC)

• Deploy models to production environments (batch and real-time inference)

• Work with distributed infrastructure (Ray / Kubernetes) for large-scale training and inference

• Monitor model performance, drift, and data quality, driving continuous improvement

• Automate retraining and continuous evaluation workflows

• Collaborate closely with Researchers, Data Engineers, and DevOps teams to enable seamless transition from research to production

• Optimize performance (CPU/GPU) and infrastructure costs

Must Have

• Proven experience as an MLOps Engineer, ML Engineer, or Data Engineer with strong ML orientation

• Strong Python skills and hands-on experience with ML/DL frameworks (PyTorch / TensorFlow)

• Experience building model training and inference pipelines

• Experience with experiment tracking and model management tools (MLflow / Weights & Biases)

• Experience with Docker and Kubernetes

• Experience working in cloud environments (AWS / GCP / Azure)

• Solid understanding of the model lifecycle and challenges of moving models to production

• Ability to work effectively in a multidisciplinary, fast-paced environment

Nice to Have

• Experience working with satellite data (EO / SAR)

• Familiarity with Ray, Airflow, or Prefect

• Experience with GPU workloads and model training optimization

• Knowledge of observability tools (Prometheus, Grafana, OpenTelemetry)

• Experience with data versioning (DVC) or feature stores

• Experience with real-time or mission-critical systems

Summary

This is a key role at the intersection of research and operational deployment, responsible for transforming models into reliable, scalable, and monitored production capabilities.

Ideal for someone looking to combine AI, data, and infrastructure while working on complex systems with real-world impact.