Senior Machine Learning Engineer

Key Skills

SQL, Python Programming, R Programming, AWS, Tableau, Scala Programming

Job Description

“We make magic.” That’s our motto at Walt Disney Parks and Resorts. And it permeates everything we do. At Disney, you’ll help inspire that magic by enabling our teams to push the limits of entertainment and create the never-before-seen!
The Lead Machine Learning Platform Engineer spearheads development and operationalization of key framework components required in the end-to-end machine learning lifecycle. Using these building blocks, Parks Data and Machine Learning Platform (“the platform”) supports a wide variety of applications in computer vision, natural language processing, data analytics, predictions, recommendations, content creation and optimization. The platform enables multiple studios of developers, data engineers, data scientists, content creators and business analysts across the entire Disney Parks Experiences and Products (DPEP) organization to build, deploy, sustain and continuously improve AI services.

Responsibilities:

• Lead a team of engineers to design and develop production grade frameworks for feature engineering, model architecture selection, model training, model interpretability, A/B testing, containerization, deployment and feedback.
• Introduce MLOps tools in conjunction with the existing DataOps and DevOps tools in the platform to facilitate collaboration across diverse user groups within the organization.
• Create automated self-service procedures that allow the platform to scale across multiple agile projects with minimal overhead.
• Define SLAs for provisioning and supporting the ML platform services.
• Integrate tools for optimizing human resource allocation in tasks involving varying degrees of complexity such as labeling, modelling, puppeting.
• Enforce effective cost optimization techniques in cloud, on-prem, and edge environments to minimize the total cost of ownership of the AI services.
• Integrate tools for measuring model KPIs, data deviation and drift.
• Provide bootstrapping templates for the frequently used ML algorithms and pipelines.
• Build reference implementations to demonstrate use of the platform.
• Guide delivery studios with accuracy-cost-explainability tradeoffs; create tools to measure them.
• Create an environment of accountable, auditable, reproducible and ethical AI by providing tools and guidance to business partners.
• Bring traditional software development best practices (unit/integration testing, documentation, service decoupling) to Machine Learning ops discipline
• Continually scan the industry and academia for new tools and methods in machine learning. Assist in defining the roadmap and evolution of the platform.
• Participate in formal/informal discussion forums to simplify and evangelize AI/ML.

Basic Qualifications:

• 3+ years of experience in training, tuning, deploying, and operationalizing ML models.
• 5+ years of experience in software development in an agile environment
• 3+ years of experience in Python
• 2+ years of experience in developing reusable frameworks.
• Expert level in software development best practices (unit/integration testing, CI/CD, service decoupling, component re-usability)
• Expert level knowledge of deep learning frameworks such as TensorFlow, MXNet, Pytorch etc.
• Expert in native AWS big data stack - S3, Glue, EMR, Sagemaker, Kinesis, Lambda, DynamoDB etc.
• Experience using infrastructure-as-a-code tools such as Cloudformation or Terraform
• Working knowledge of ML model architectures and hyper-parameter tuning.
• Working knowledge of libraries such as scikit-learn, keras, pandas, svm, xgboost etc.
• Hands-on experience with ML lifecycle tools such as AWS Sagemaker, MLflow, TensorFlow Serve etc.
• Hands-on experience with coding, troubleshooting and performance tuning ML pipeline involving large datasets in a collaborative project environment.
• Working knowledge of Apache Spark
• Experience in using 3rd party computer vision algorithms (e.g OpenCV) and video processing libraries (e.g. NVIDIA DeepStream)
• Knowledge of various media formats, encoding schemes and streaming protocols for computer vision ML pipeline.
• Knowledge of various relational/non-relational databases and storage technologies.
• Knowledge of high performance databases and caching solutions (e.g. Redis)
• Experience in accelerating deep learning frameworks using GPU and micro-batching
• Familiarity with ongoing research in model explainability; theoretical knowledge of techniques such as LRP, CEM, LIME.

Required Education

• Master’s Degree in Computer Science, Distributed Computing, Machine Learning, Mathematics or related field, or equivalent work experience