Senior Machine Learning Engineer

We are seeking a highly skilled and experienced Senior Machine Learning Engineer to drive the development, deployment, and monitoring of cutting-edge machine learning solutions in a scalable cloud environment. You will collaborate with a cross-functional team to deliver end-to-end ML pipelines, ensure high-quality model performance, and implement robust monitoring for production systems.

Responsibilities

• Reproduce training environments with pinned dependencies and seeds, rebuild feature and label pipelines, and register models
• Deploy SageMaker endpoints and batch jobs with autoscaling, execute shadow/canary strategies, and define rollback protocols
• Set up ML experiments, Model Registry, and Feature Stores, integrating them into CI/CD workflows
• Implement Model Monitor for drift and performance tracking, configure alerting mechanisms, and create comprehensive runbooks
• Optimize pipelines for reliable model serving, ensuring low latency, high quality, and rollback readiness during hypercare
• Collaborate with data engineers to build scalable data pipelines on S3/Snowflake for training and inference workflows

Requirements

• 3+ years of hands-on experience in machine learning engineering, with a solid background in Python 3.x
• Proficiency in PyTorch or TensorFlow, with a track record of building and deploying models in production
• Expertise in AWS SageMaker, including Pipelines, Registry, Endpoints, and Model Monitor
• Familiarity with MLflow or SageMaker Experiments for experiment tracking and artifact management
• Strong skills in Docker for containerization of ML workloads
• Experience with Feature Store solutions such as SageMaker Feature Store or Feast
• English level B1+ for effective communication

Nice to have

• Proficiency in FastAPI or BentoML for lightweight API service development
• Knowledge of ONNX for model optimization and cross-platform interoperability
• Familiarity with distributed inference tools like Ray Serve or Triton
• Experience with large-scale distributed training using Horovod or DeepSpeed
• Understanding of Hugging Face Transformers for state-of-the-art NLP model implementation