Prantik Howlader is an Applied Scientist II and PhD researcher with nine years of industry and academic experience building deep learning and multimodal AI systems. His work spans agentic AI and reinforcement learning for tool use, multimodal segmentation with VLMs, and practical deployment of lightweight Transformer-based segmentation on constrained devices. He has driven applied research at Amazon and Zebra Technologies—improving product recommendations and customer return experiences via context-aware visual question generation and multimodal agents—and continues PhD research on low-label, class-imbalanced segmentation for medical and natural images. Early roles in networking and systems at Cisco and Wipro give him a rare combination of production-grade software engineering and cutting-edge CV/ML research. Notably, he has applied VAE data augmentation and token-efficient detector-to-transformer pipelines to tackle class imbalance and resource limits in real deployments.
9 years of coding experience
7 years of employment as a software developer
Don Bosco School, Liluah
Bachelor of Technology (B.Tech.) Information Technology, Bachelor of Technology (B.Tech.) Information Technology at West Bengal University of Technology, Kolkata
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.