• 6–8 Years of Engineering: Proven track record building and operating production backend services at scale.
• Expertise: Deep mastery of JVM languages (Kotlin preferred, Java acceptable), Python(FastAPI), distributed systems patterns, and cloud-native service design (Kubernetes, Azure, GCP, AWS, Private cloud).
• Workflow & Orchestration: Hands-on experience with workflow engines (Conductor, Apache Airflow, or equivalent) and event-driven architectures (Kafka, Pub/Sub).
• Data & Storage: Comfortable working with relational databases (PostgreSQL) & non-relational databases, object storage(Data-lakes), and caching layers (Redis) in multi-tenant environments.
• Observability Stack: Practical experience with Open-telemetry, Prometheus, and Grafana for instrumentation and operational insight.
• ML Platform Exposure: experience supporting ML workloads & notebooks in production, whether through job scheduling, resource management, experiment tracking integration, or model serving infrastructure.
• Contextualisation Domain (Bonus): Familiarity with industrial knowledge graph construction, entity resolution, or NLP/CV pipelines as they relate to industrial asset data is a strong differentiator.
• Full-Stack Awareness (Bonus): Familiarity with React or TypeScript is a plus for consuming and dogfooding your own platform’s developer tooling.
• The Platform Thinking Spirit: A passion for building composable, well-documented, and automated platform systems that empower other engineers including ML engineers to build faster.

Good to have

ML Platform & Contextualisation:

• ML Workload Support: Build and extend platform primitives compute scheduling, environment management, and secrets handling, that enable ML engineers to run model training, fine-tuning, and batch inference jobs reliably.
• Contextualisation Pipelines: Support the engineering infrastructure behind Cognite’s Contextualisation capabilities (entity matching, asset hierarchy inference, P&ID parsing) by ensuring the platform can orchestrate long-running, GPU-aware, and data-intensive ML workflows without manual intervention.
• Vector & Embedding Infrastructure (Bonus): Familiarity with serving or storing vector embeddings to support semantic search and RAG-based contextualisation use cases.
• Model Lifecycle Awareness: Understand model versioning, A/B experiment tracking, and the boundary between platform concerns and ML framework concerns, so the platform stays lean while ML teams stay unblocked.