Research Scientist - Structured data modeling

About the Company

Data Intelligence Lab researches and develops cutting-edge Deep Learning algorithms to perform advanced data analysis, prediction, and complex data optimization. The OO Tech Cell aims to develop proprietary Tabular Foundation Models that are generalizable, explainable, and scalable across various industries.

Responsibilities

• Analyze and improve the structure, training methods, and inference mechanisms of Tabular Foundation Models
• Evaluate performance on diverse tabular datasets and derive improvement hypotheses through failure case analysis
• Research explainability and interpretability functions such as feature importance and SHAP
• Study methods to combine LLMs and multimodal models with tabular models
• Document research results and collaborate with engineering teams to implement findings into APIs and products

Requirements

• Master's degree with 2+ years of experience or a PhD in a related field
• Ability to understand and propose improvements for deep learning model architectures
• Understanding of Transformer, attention, representation learning, and self-supervised learning
• Proficiency in reading, designing, and implementing research code using PyTorch
• Strong interest in tabular or structured data problems
• Ability to analyze experimental results and failure cases to form new research hypotheses
• Ability to quickly grasp core ideas from papers, technical documents, and open-source implementations

Preferred Qualifications

• Research or implementation experience in Tabular learning, AutoML, TabPFN, or in-context learning
• Experience with Foundation models, meta-learning, few-shot learning, or representation learning
• Research experience in Explainable AI (XAI) or interpretable ML
• Experience combining LLM/multimodal models with tabular data
• Experience writing and submitting papers to academic conferences
• Basic understanding of model serving, API development, and MLOps