MLOps Engineer - Model Deployment

Job Description :

We are looking for a highly motivated individual who can develop cutting edge MLOps and DevOps frameworks to deploy AI models. The candidate should have a solid grasp of state-of-the-art cloud technologies, best in class deployment architectures / frameworks and production grade software.

The candidate must have an open mindset regarding the tradeoffs between coding from scratch versus utilizing existing frameworks. Many of the problems we encounter are novel and have never been solved before, so creative, out-of-the-box thinking and a fondness for experimentation are a must. We also want someone who stays current with recent trends in MLOps / DevOps so our approaches remain the most robust and competitive in the industry. Finally, the role requires strong team and interdisciplinary collaboration to see products through the development cycle from beginning to end.

Core Job Responsibilities :

• Develop end-to-end ML pipelines encompassing the ML lifecycle from data ingestion, data transformation, model training, model validation, model serving, and model evaluation over time.
• Collaborate closely with AI scientists to accelerate productionization of ML Setup CI / CD / CT pipelines for ML Deploy models as a service both on-cloud and on-prem.
• Learn and apply new tools, technologies, and industry best Qualifications :
• MS in Computer Science, Software Engineering, or equivalent field
• Experience with Cloud Platforms - GCP and Azure, and related skills : Docker, Kubernetes, edge computing
• Familiarity with task orchestration tools such as MLflow, Kubeflow, Airflow, Vertex AI, Azure ML, etc.
• Fluency in at least one general purpose programming language.
• Python or Java, Strong DevOps skills : Linux / Unix environment, testing, troubleshooting, automation, Git, ,dependency management, and build tools (GCP Cloud Build, Jenkins, Gitlab CI / CD, Github Actions, etc.).
• Data engineering skills a plus, such as Beam, Spark, Pandas, SQL, Kafka, GCP Dataflow etc.

ref : hirist.tech)