Machine Learning Engineer - Document Intelligence

Workday Workday · Enterprise · Pleasanton, CA +1

Machine Learning Engineer focused on building and optimizing Workday's Document Intelligence Platform. This involves designing and implementing LLM-based technologies for document parsing, entity extraction, and classification, as well as traditional ML techniques. The role includes building scalable ML pipelines for data preprocessing, training, and inference, and working with RAG architectures and agentic frameworks for document intelligence applications.

What you'd actually do

  1. Support the design and implementation of LLM-based technologies for document parsing, entity extraction, and classification tasks.
  2. Apply traditional ML and deep learning techniques to continuously enhance the accuracy, efficiency, and scalability of our document intelligence models.
  3. Build scalable ML pipelines and services for data preprocessing, feature engineering, training, and inference, enabling high-volume document processing workflows.
  4. Perform exploratory data analysis (EDA) on diverse document datasets to uncover valuable insights, optimize feature engineering, and inform model development.

Skills

Required

  • 3+ years of experience researching, developing and deploying production-grade ML systems
  • expertise in deep learning, NLP, Information Retrieval, and recommender systems using frameworks like PyTorch or TensorFlow
  • Proven track record of building and evaluating NLP and LLM-powered products
  • expertise in RAG architectures, agentic frameworks (e.g., LangChain/LangGraph), and long-context LLM applications (e.g., Text-to-SQL)
  • 2+ years of Python experience with a focus on modular library design, asynchronous patterns, and scalable system architecture (state management/error handling) for non-deterministic AI outputs

Nice to have

  • Advanced degree (Master’s or Ph.D.) in a quantitative field or a strong portfolio of peer-reviewed research publications
  • Proficiency in techniques like DSPy, Reinforcement Learning, imitation learning, graph neural networks, multi-modal models, and large-scale data processing (PySpark, SQL)
  • A "test-everything" mindset with experience in A/B testing, Knowledge Graphs, and "Golden Dataset" curation for model benchmarking
  • Proficiency in large-scale data processing (PySpark, SQL)
  • Hands-on experience with the full ML lifecycle, including model fine-tuning (PEFT), evaluation frameworks (e.g., DeepEval/RAGAS), and cloud-native deployment (Docker/K8s, AWS/GCP)
  • Demonstrated ability to lead cross-functional teams, mentor junior engineers, and solve ambiguous problems with high autonomy

What the JD emphasized

  • production-grade ML systems
  • deep learning
  • NLP
  • Information Retrieval
  • recommender systems
  • Generative AI & Agentic Systems
  • NLP and LLM-powered products
  • RAG architectures
  • agentic frameworks
  • long-context LLM applications
  • Python experience
  • scalable system architecture
  • non-deterministic AI outputs
  • model fine-tuning
  • evaluation frameworks

Other signals

  • LLM-based technologies
  • document parsing
  • entity extraction
  • classification tasks
  • NLP
  • computer vision
  • large language models (LLMs)
  • in-house model training
  • entity resolution
  • document processing pipelines
  • RAG architectures
  • agentic frameworks
  • long-context LLM applications