onsite
Data Scientist - Visual Document Understanding
Data Scientist
Data Scientist specialized in visual document understanding, responsible for designing complex evaluation tasks, defining ground‑truth outputs, and creating objective rubrics to assess AI models that follow instructions on document images.
About the role
Key Responsibilities
- Design and author sophisticated evaluation tasks for AI models that process and understand visual documents.
- Define clear, unambiguous ground‑truth outputs for each task to ensure reliable benchmarking.
- Develop objective, reproducible rubrics and scoring criteria to measure model performance on instruction‑following and document comprehension.
- Collaborate with model engineers to integrate evaluation pipelines into continuous testing frameworks.
- Analyze results, identify failure modes, and provide actionable feedback for model improvement.
Requirements
- Strong background in Data Science with hands‑on experience in Machine Learning and Computer Vision.
- Proficiency in Python and related libraries (e.g., PyTorch, TensorFlow, OpenCV).
- Experience creating high‑quality data annotations, ground‑truth datasets, and evaluation rubrics.
- Familiarity with prompt engineering and instruction‑following model assessment.
- Excellent analytical and communication skills to translate evaluation findings into clear recommendations.
Skills
pythonmachine learningcomputer vision