Diffusion models for time-series (forecasting, anomaly detection, imputation, classification)

Worked on Python SDK, image generation, and fine-tuning diffusion models.

Multimodal contrastive learning for hate speech detection; KD & diffusion for image segmentation/classification.

Data quality & synthetic data generation for Indic languages

Automated data extraction pipelines for financial data(2M+ records, spanning over 15 years)