About the role
Key Responsibilities
ML Infrastructure & Tooling
• Set up and own the ML infrastructure on the data platform — model registry, experiment tracking, feature store, and serving infrastructure.
• Define and enforce MLOps standards — how models are versioned, how experiments are tracked, and what production readiness criteria must be met before deployment.
• Build and maintain the CI/CD pipelines for model development and deployment, enabling the team to ship model updates reliably and repeatedly.
Model Deployment & Serving
• Own the deployment of all models to production — from the ML Engineer’s trained model to a stable, monitored, serving endpoint.
• Design and implement model serving infrastructure capable of handling real-time inference requirements with appropriate latency and throughput.
• Build fallback mechanisms for all deployed models — when model confidence falls below defined thresholds, the system must degrade gracefully to rule-based logic rather than produce unreliable outputs.
• Manage model versioning and rollback capabilities so that problematic deployments can be reversed quickly and safely.
Monitoring & Drift Detection
• Build and maintain model performance monitoring dashboards covering accuracy, latency, throughput, and data drift indicators.
• Define performance thresholds for each deployed model and implement alerting when models deviate from expected behaviour.
• Build automated retraining pipelines triggered by drift detection or performance degradation, ensuring models stay current as data and regulations evolve.
• Maintain a feedback loop between validator decisions in the application layer and model retraining workflows.
Production Reliability
• Own the operational readiness of the AI system for production deployment — monitoring, alerting, runbooks, and incident response procedures.
• Work with the Backend Engineer to ensure model inference endpoints meet the latency and reliability requirements of the integrations that depend on them.
• Conduct regular load testing and capacity planning to ensure the serving infrastructure scales with growing transaction volumes.
Required Qualifications
Education
• B.E. / B.Tech / M.Tech in Computer Science, Information Technology, or a related field.
Experience
• 5+ years of engineering experience with at least 3 years focused on MLOps or ML infrastructure in a production environment.
• Proven experience deploying and maintaining ML models at production scale — not research or experimentation contexts.
• Experience building model monitoring and automated retraining pipelines.
Technical Skills
• ML Platforms: MLflow for experiment tracking and model registry; Databricks MLOps stack strongly preferred.
• Model Serving: Experience with model serving frameworks — TorchServe, TensorFlow Serving, BentoML, or equivalent; Databricks Model Serving a plus.
• Infrastructure: Docker and Kubernetes for containerised model deployment; Terraform or equivalent for infrastructure-as-code.
• CI/CD: GitHub Actions, GitLab CI, or equivalent for automated model deployment pipelines.
• Monitoring: Prometheus, Grafana, or equivalent for infrastructure and model performance monitoring; experience with data drift detection frameworks.
• Languages: Python proficiency; familiarity with Bash scripting for automation.
• Cloud: Experience with at least one major cloud provider — AWS, Azure, or GCP — for managed ML infrastructure services.
Preferred Qualifications
• Experience operating ML systems with real-time inference requirements in an enterprise setting.
• Familiarity with feature stores and their role in maintaining consistency between training and serving environments.
• Prior work in fintech, compliance, or similarly regulated environments where model explainability and auditability are requirements.
• Experience with A/B testing and shadow deployment patterns for safe model rollouts.
• Contributions to MLOps open-source tooling or community knowledge sharing.