Senior Machine Learning Engineer, Synthetic Data & Document Understanding

ABBYY ABBYY · Enterprise · India · R&D (Engineering)

This role focuses on building generative pipelines for synthetic data at scale for document understanding tasks. The engineer will design and implement systems to produce high-quality, diverse synthetic training data, develop evaluation frameworks, and ensure the synthetic data improves downstream model performance. Responsibilities include owning the synthetic data generation track end-to-end, driving architectural decisions, and building scalable, cost-efficient pipelines.

What you'd actually do

  1. Design and implement pipelines that analyze real documents to inform high-fidelity synthetic data generation
  2. Build generative systems capable of producing documents across diverse formats, layouts, and domains
  3. Develop evaluation frameworks to ensure synthetic data maintains distributional fidelity and diversity
  4. Own the synthetic data generation track end-to-end, from architecture to quality validation
  5. Build scalable pipelines capable of generating millions of synthetic training examples

Skills

Required

  • MS or PhD in Computer Science, Engineering, Mathematics, or related field
  • 5+ years of experience in Machine Learning / AI
  • Generative models
  • Vision-Language Models (VLMs)
  • Synthetic data systems
  • Python
  • PyTorch or similar frameworks
  • Data quality evaluation
  • Statistical analysis
  • Large-scale data pipelines
  • Cloud environments
  • Experiment tracking

Nice to have

  • document understanding (layout, structure, semantics)
  • realistic noise patterns
  • domain coverage

What the JD emphasized

  • Proven experience building and evaluating synthetic data pipelines for ML training
  • Deep expertise in Vision-Language Models and document understanding
  • Strong knowledge of generative modeling for structured and semi-structured data
  • Experience evaluating data quality via automated metrics and downstream model impact

Other signals

  • synthetic data generation
  • generative models
  • document understanding
  • ML data engineering