Machine Language Engineer

Description:

We are looking for competence in designing, configuring, deploying, and maintaining machine learning solutions for anomaly detection and early warning in a data center environment. The solution must run on-premises and integrate with existing operational data streams, including InfluxDB as the time-series database and Kafka as part of the data pipeline. The role requires practical experience with machine learning, time-series analytics, Linux operations, and production deployment in industrial or infrastructure environments.

Required Competence

• Experience with time-series data analysis and anomaly detection for infrastructure systems such as cooling, power, UPS, environmental sensors, network equipment, and IT load.
• Ability to evaluate and select suitable ML models for anomaly detection, forecasting, and early warning use cases.
• Knowledge of statistical anomaly detection, Isolation Forest, Autoencoders, LSTM models, Prophet, ARIMA, streaming anomaly detection, and hybrid rule/ML-based systems.
• Ability to explain trade-offs related to accuracy, explainability, training data requirements, operational complexity, false positives, and real-time performance.

Programming and Development Competence

• Experience with recognized programming languages commonly used for machine learning and data engineering such as Python, Go, Java, Scala, Rust, or C++.
• Understanding of frameworks and tooling for machine learning pipelines, data processing, and model serving.
• Experience building APIs and backend services for operational environments.
• Ability to work with containerized and distributed systems.

Technical Environment

• Ubuntu Linux configuration, hardening, and operations.
• InfluxDB integration for time-series storage and retrieval.
• Kafka integration for streaming pipelines and event-driven architectures.
• Docker or Kubernetes-based deployment.
• REST API development using frameworks such as FastAPI, Spring Boot, Go services, or similar technologies.
• Logging, monitoring, alerting, and production operations.

Data Pipeline and Integration

• Reading real-time or near-real-time data from Kafka.
• Querying historical data from InfluxDB for model training.
• Writing anomaly scores and prediction results back to InfluxDB.
• Exposing results through APIs, dashboards, or Grafana integrations.
• Supporting integration with ITSM or operational monitoring systems.

Deployment Responsibilities

• Installing required dependencies and runtime environments.
• Configuring services using systemd, Docker, or Kubernetes.
• Managing model artifacts and versioning.
• Scheduling retraining and inference jobs.
• Implementing health checks, logging, and operational metrics.
• Documenting deployment, rollback, and maintenance procedures.

Operational Requirements

• Designing robust and explainable anomaly detection systems.
• Reducing false positives while ensuring meaningful early warning capability.
• Defining training data requirements and retention periods.
• Handling missing data, sensor errors, and seasonal operational patterns.
• Distinguishing between normal operational variation and real anomalies.

Desired Deliverables

• Recommended ML model architecture.
• Data pipeline design using Kafka and InfluxDB.
• Ubuntu deployment guide.
• Configuration files and service definitions.
• Model training and inference implementation.
• Alerting and anomaly scoring logic.
• Operational and maintenance documentation.