Principal Software Engineer (workflow / platform)
Cognite - AI for IndustryResponsibilities
- Design, build, and operate the core serverless execution engine and Workflows orchestration layer for Cognite Data Fusion’s AI and automation capabilities.
- Own uptime, latency SLOs, incident response, deterministic function execution, and reliable workflow progression.
- Architect multi-tenant and multi-cloud systems with high-throughput scheduling, queueing, retry, and graceful-degradation mechanisms.
- Define and evolve versioned REST and event-driven APIs for internal engineering teams and external customers.
- Implement distributed tracing, structured logging, alerting, and operational observability.
- Maintain test automation and deployment pipelines for unit, integration, and smoke testing.
- Profile and optimize execution throughput, cold-start latency, cross-service calls, and cloud compute and storage costs.
- Support ML workloads, including compute scheduling, environment management, secrets handling, model training, fine-tuning, batch inference, and long-running data-intensive workflows.
Requirements
- 10+ years of experience building and operating production backend services at scale.
- Deep expertise in JVM languages, preferably Kotlin or alternatively Java, plus Python and FastAPI.
- Strong knowledge of distributed systems patterns and cloud-native service design.
- Experience with Kubernetes and cloud environments including Azure, GCP, AWS, and private cloud.
- Hands-on experience with workflow engines such as Conductor or Apache Airflow and event-driven architectures using Kafka or Pub/Sub.
- Experience with relational and non-relational databases, object storage and data lakes, and caching layers such as Redis in multi-tenant environments.
- Practical experience with OpenTelemetry, Prometheus, and Grafana for instrumentation and operational insight.
- Experience supporting production ML workloads, notebooks, job scheduling, resource management, experiment tracking, or model-serving infrastructure.
- Familiarity with industrial knowledge graphs, entity resolution, NLP or computer vision pipelines, React, or TypeScript is beneficial.
- Familiarity with vector embeddings, semantic search, retrieval-augmented generation, model versioning, and A/B experiment tracking is beneficial.
Tech Stack
About Cognite - AI for Industry
Cognite is the only Industrial Data and AI platform built for scaling. We help Energy, Manufacturing, and Power & Utilities companies break down data silos and turn complex operational data into actionable, enterprise-wide value. Why Cognite? Unify: Cognite Data Fusion® unifies OT, ET, and IT data into a real-time Industrial Knowledge Graph. Automate: Cognite Atlas AI™ deploys low-code Industrial AI Agents to supercharge SME capacity. Deliver(2025): We drive a 400% ROI (Forrester) and have identified over $1B in customer value in the past year alone. Our Moonshot We are on a mission to deliver $100 billion in realized customer value by 2035. From increasing production uptime to reducing downtime by 30%, Cognite provides the foundation to scale AI solutions across every asset and site. Founded in 2016, we now number 700 strong, including some of the best software developers, data scientists, designers, and 3D specialists in the field. Get to know us better: Make data do more: makedatadomore.cognite.com/ Twitter: @CogniteData Facebook: @CogniteData IG: @CogniteData Youtube: /c/cognite