Principal Software Developer- MLOps Platform

Job Requisition ID #

25WD87574

25WD87574, Principal Software Developer- MLOps Platform

French translation to follow!/Traduction française à suivre!

Position Overview

We are looking for an experienced Principal Software Engineer to join our platform team focusing on AI/ML Platform (AMP). This team builds and maintains central components to fast track the development of new ML/AI models such as model development studio, feature store, model serving, and model observability. The ideal candidate would have a background in MLOps, Data engineering, and DevOps with experience in building high-scale deployment architectures and observability. As an important contributor to our engineering team, you will help shape the future of our AI/ML capabilities, delivering solutions that inspire value for our organization. You will report to a manager.

Responsibilities

• System design: You will design, implement and manage software systems for the AI/ML Platform, orchestrating the full ML development lifecycle for the partner teams.

• Mentoring: Spread your knowledge, share best practices, and conduct design reviews to enhance expertise at the team level.

• Multi-cloud architecture: Define components that leverage strengths from multiple cloud platforms (e.g., AWS, Azure) to optimize performance, cost, and scalability.

• AI/ML observability: Build systems for monitoring the performance of AI/ML models and extracting insights on the underlying data such as drift detection, data fairness/bias, and anomalies.

• ML Solution Deployment: Develop tools for building and deploying ML artifacts in production environments, facilitating a smooth transition from development to deployment.

• Big Data Management: Automate and orchestrate tasks related to managing big data transformation and processing, building large-scale data stores for ML artifacts.

• Scalable Services: Design and implement low-latency, scalable prediction and inference services to support the diverse needs of our users.

• Cross-Functional Collaboration: Collaborate across diverse teams, including machine learning researchers, developers, product managers, software architects, and operations, fostering a collaborative and cohesive work environment.

• End-to-end ownership: Take end-to-end ownership of the components and work with other engineers in the team including design, architecture, implementation, rollout, onboarding support to partner teams, production on-call support, testing/verification, and investigations.

Minimum Qualifications

• Educational Background: Bachelor's degree in Computer Science or equivalent practical experience.

• Experience: Over 8 years of experience in software development and engineering, delivering production systems and services.

• Prior experience of working with MLOps teams at the intersection of expertise across ML model deployments, DevOps, and data engineering.

• Hands-on skills: Ability to fluently translate the design into high-quality code in Golang, Python, and Java.

• Knowledge of DevOps practices, containerization, and orchestration tools such as CI/CD, Terraform, Docker, Kubernetes, and GitOps.

• Demonstrated knowledge of distributed data processing frameworks, orchestrators, and data lake architectures using technologies such as Spark, Airflow, and iceberg/parquet formats.

• Prior collaborations with Data science teams to deploy their models, setting up ML observability for inference-level monitoring.

• Exposure to building RAG based applications by collaborating with other product teams, Data scientists/AI engineers.

• Demonstrated creative problem-solving skills with the ability to break down problems into manageable components.

• Knowledge of Amazon AWS and/or Azure cloud for solutioning large scale application deployments.

• Excellent communication and collaboration skills, fostering teamwork and effective information exchange.

Preferred Qualifications

• Experience of integrating with third-party vendors.

• Experience in latency optimization with the ability to diagnose, tune, and enhance the efficiency of serving systems.

• Familiarity with tools and frameworks for monitoring and managing the performance of AI/ML models in production (e.g., MLflow, Kubeflow, TensorBoard).

• Familiarity with distributed model training/inference pipelines using KubeRay or equivalent.

• Exposure to leveraging GPU computing for AI/ML workloads, including experience with CUDA, OpenCL, or other GPU programming tools, to significantly enhance model training and inference performance.

• Exposure to ML libraries such as PyTorch, TensorFlow, XGBoost, Pandas, and Scikit-Learn.