MLOps Engineer

EPAM Systems Inc; Bengaluru; In loco; INR 4.000.000 - 6.000.000; Tempo pieno

Descrizione del lavoro

EPAM Systems Inc. is seeking an experienced MLOps Tech Lead to own the end-to-end ML lifecycle across cloud and on-prem environments.

You will build automated pipelines, integrate labeling workflows, and collaborate with AI/Framework/Data engineering teams to ensure scalable, production-ready ML systems for autonomous driving. You will lead efforts in training automation, model optimization, deployment to cloud and embedded platforms, and continuous monitoring.

Competenze

• Bachelor's or master's degree in engineering, computer science, robotics, or a related field.
• 5+ years (tech lead) of experience in data labeling, data operations, data quality, or ML dataset management, preferably in autonomous driving.
• Strong programming skills in Python, with familiarity in C++ preferred.
• Hands-on experience with MLOps frameworks: MLflow, Kubeflow, Airflow, Argo, Azure ML, Databricks, or similar.
• Strong understanding of Docker, Kubernetes, CI/CD, and distributed training.
• Experience with data pipeline technologies (Kafka, Spark, Delta Lake, Databricks).
• Good understanding of multimodal autonomous driving data (camera, LiDAR, radar).
• Proven experience deploying models to cloud, edge, or embedded devices.
• Experience with real-time systems and embedded CI/CD pipelines.

Mansioni

• Build and maintain end-to-end ML pipelines across cloud and on-prem environments.
• Ensure ML pipelines run in cloud and on-prem GPU clusters with reproducibility and traceability.
• Integrate data pipelines with labeling platforms and automate dataset creation and quality checks.
• Work with the label manager to enforce labeling quality gates and track dataset KPIs.
• Automate training workflows, hyperparameter tuning, experiment tracking, and benchmarking.
• Integrate optimization steps (quantization, compilation, pruning, distillation) with the framework team.
• Establish deployment pipelines for models to cloud platforms, on-prem HPC/GPU clusters, and embedded compute platforms.
• Develop automated monitoring for model drift, data drift, performance degradation, anomalies, and operational metrics.
• Build dashboards and alerts to monitor model performance across simulations and real-world tests.
• Drive MLOps best practices, documentation, and standardized ML workflows; steer MLOPS engineers.

Conoscenze

Python

C++

MLOps leadership

Data labeling

Cloud & on-prem deployment

Formazione

Engineering or CS degree

Strumenti

MLflow

Kubeflow

Airflow

Argo

Azure ML

Databricks

Docker

Kubernetes

Kafka

Spark

Delta Lake

Descrizione del lavoro

MLOps is responsible for building and maintaining the full machine learning lifecycle across cloud and on-prem environments. This includes integrating data pipelines, labeling flows, training pipelines, testing frameworks, model optimization, deployment to cloud and embedded hardware, and monitoring of model performance. The role works closely with AI engineers, framework engineers, application engineers, data engineering, and label management to enable scalable, reliable, and production-ready ML systems for autonomous driving.

The core responsibilities for the job include the following:

End-to-End MLOps Pipeline Development:

• Build and maintain automated end-to-end ML pipelines covering data ingestion, dataset management, labeling workflows, training, validation, optimization, deployment, and monitoring.
• Ensure ML pipelines run in both cloud and on-prem GPU clusters, with strong reproducibility and traceability.

Data and Labeling Integration:

• Integrate data pipelines with labeling platforms and automate dataset creation and quality checks.
• Work closely with the label manager to enforce labeling quality gates and track dataset KPIs.

Training, Testing, and Optimization:

• Support AI Engineers by automating training workflows, hyperparameter tuning, experiment tracking, and benchmarking.
• Integrate optimization steps (quantization, compilation, pruning, distillation) in collaboration with the Framework Engineer.

Model Deployment:

• Establish deployment pipelines for ML models to cloud platforms (Azure/AWS), on-prem HPC/GPU clusters (Kubernetes, Slurm, and NVIDIA infrastructure), and embedded compute platforms (NVIDIA Orin).

Monitoring and Observability:

• Develop automated monitoring systems for model drift, data drift, performance degradation, anomalies, and operational metrics.
• Build dashboards and alerts to monitor model performance across simulation and on-vehicle tests.

Collaboration and Cross-Functional Alignment:

• Work closely with AI engineers in training and evaluation, framework engineers in model optimization, application engineers integrating ML outputs into the autonomy stack, data engineers on scalable data and storage architecture, and ML architects on integration and implementation of new components.
• Drive best practices for MLOps, documentation, and standardized ML workflows across the department.
• Steer and guide MLOPS engineers.

Requirements:

• Bachelor's or master's degree in engineering, computer science, robotics, or a related field.
• 5+ years (tech lead) of experience in data labeling, data operations, data quality, or ML dataset management, preferably in autonomous driving.
• Strong programming skills in Python, with familiarity in C++ preferred.
• Hands-on experience with MLOps frameworks: MLflow, Kubeflow, Airflow, Argo, Azure ML, Databricks, or similar.
• Strong understanding of Docker, Kubernetes, CI/CD, and distributed training.
• Experience with data pipeline technologies (Kafka, Spark, Delta Lake, Databricks).
• Good understanding of multimodal autonomous driving data (camera, LiDAR, radar).
• Proven experience deploying models to cloud, edge, or embedded devices.
• Experience with real-time systems and embedded CI/CD pipelines.