Siddharth Choudhary I am a Research Scientist at Luma AI. Before this, I was a Principal Applied Scientist in Amazon AGI, where I was a tech-lead on multimodal pre-training team for the Amazon Nova family of foundation models. My work focuses on developing next-generation AI systems that seamlessly integrate vision, speech, and language understanding with generative capabilities. My research spans the full spectrum of AI and computer vision—from foundation models and vision-language systems to robotics and SLAM. I've published at top-tier venues including CVPR, IJRR, ICRA, and Nature Digital Medicine, with over 1,200 citations and an h-index of 14. My work on multimodal hallucination control was featured at CVPR 2024 and highlighted in AWS's keynote presentation. Before Amazon, I was a Principal Computer Vision Engineer at Magic Leap, where I architected the 3D object recognition system deployed in Magic Leap One. I earned my Ph.D. in Computer Science from Georgia Tech, focusing on distributed algorithms for multi-robot SLAM systems. I'm passionate about pushing the boundaries of multimodal LLMs and building systems that bridge the gap between understanding and generation. Checkout itzsid.gith
Research Scientist
Principal Computer Vision Researcher/Engineer
Master's studies in software
Bachelor's studies in software
