An Assistant Research Professor affiliated with the Laboratory for Computational Sensing and Robotics at Johns Hopkins University.
How media typically covers Manish Sahu
Manish Sahu as author
Grab engineered a lightweight, specialized Vision LLM based on Qwen2-VL 2B to accurately extract information from diverse Southeast Asian document formats, addressing limitations of traditional OCR and proprietary LLMs.
“Co-author of the article on building a custom Vision LLM for document processing at Grab”