MLOps Platform & Data Pipeline Engineer

Build secure, sovereign pipelines for massive multi-spectral training imagery: storage clusters, automated labeling, and high-frequency training and evaluation.

Good models need clean, fast, secure data pipelines, and ours carry sensitive multi-spectral imagery that can never leak. You will build the sovereign infrastructure — storage clusters, automated labeling, training, and continuous evaluation — that turns field capture into deployable edge models. Expect Kubernetes, GPU clusters, and a zero-trust mindset on every access path. You are the reason a model improvement found on Monday can be re-validated and shipped by the end of the week.

• Docker, Kubernetes, object storage, and high-scale relational systems
• MLOps pipelines (Kubeflow/MLflow) and labeling tools (CVAT/Label Studio)
• Python, PyTorch, GPU drivers, and Linux clustering
• Zero-trust access control discipline

• Maintain on-prem training infrastructure and clusters
• Scale video ingestion and chunking for new tactical logs
• Run continuous evaluation benchmarks for edge models
• Secure model artifacts with signature validation