Seamless deployment, monitoring, and optimization of AI models in production
Requirements
- Proficiency in Python and SQL; familiarity with JavaScript or Go is a plus.
- Expertise in containerization (Docker, Kubernetes) and CI/CD tools (GitHub Actions, Jenkins).
- Knowledge of time-series databases (e.g., InfluxDB, TimescaleDB) and logging frameworks (e.g., ELK Stack, OpenTelemetry).
- Experience with drift detection tools (e.g., Evidently AI, Alibi Detect) and visualization libraries (e.g., Plotly, Seaborn).
- Understanding of model performance metrics (e.g., precision, recall, AUC) and drift detection methods (e.g., KS test, PSI).
- Familiarity with AI vulnerabilities (e.g., data poisoning, adversarial attacks) and mitigation tools like Adversarial Robustness Toolbox (ART).
- Experience with MLflow, Kubeflow, or cloud platforms (AWS SageMaker, Azure ML) for deploying models in production.
Responsibilities
- Model Deployment: Deploy and manage machine learning models in production using tools like MLflow, Kubeflow, or AWS SageMaker, ensuring scalability and low latency.
- Monitoring and Observability: Build and maintain dashboards using Grafana, Prometheus, or Kibana to track real-time model health (e.g., accuracy, latency) and historical trends.
- Data Drift Detection: Implement drift detection pipelines using tools like Evidently AI or Alibi Detect to identify shifts in data distributions and trigger alerts or retraining.
- Logging and Tracing: Set up centralized logging with ELK Stack or OpenTelemetry to capture AI inference events, errors, and audit trails for debugging and compliance.
- Pipeline Automation: Develop CI/CD pipelines with GitHub Actions or Jenkins to automate model updates, testing, and deployment.
- Security and Compliance: Apply secure-by-design principles to protect data pipelines and models, using encryption, access controls, and compliance with regulations like GDPR or NIST AI RMF.
- Optimization: Optimize models for production (e.g., via quantization or pruning) and ensure efficient resource usage on cloud platforms like AWS, Azure, or Google Cloud.
Other
- Bachelor’s or Master’s degree in Computer Science, Data Science, Engineering, or a related field.
- 5+ years in MLOps, DevOps, or software engineering with a focus on AI/ML systems.
- Strong problem-solving and debugging skills for resolving pipeline and monitoring issues.
- Excellent collaboration and communication skills to work with cross-functional teams.
- Attention to detail for ensuring accurate and secure dashboard reporting.
