Advanced Model Research: Lead the R&D of high-performance models for Fine-Grained Visual Categorization (FGVC) and Hierarchical Classification, focusing on detecting complex object attributes and nuances in unconstrained environments using both Supervised and Self-Supervised Learning (SSL).

Generative AI & VLM: Spearhead the exploration of Vision Language Models (VLM) and CLIP-based architectures to enable Text-to-Image / Image-to-Image retrieval and Open-Set Recognition, moving beyond fixed-class taxonomy.

Agentic Data Workflows: Architect and implement AI Agentic Workflows to automate the data lifecycle. Design intelligent agents for autonomous data mining, auto-captioning, and Synthetic Data Generation to overcome long-tail edge cases and reduce reliance on manual annotation.

Production-Grade AI: Drive the transition of models from research prototypes to production. Collaborate with Embedded AI ...