About Pujari Keerthika
Pujari Keerthika (also written Keerthika Pujari) is an undergraduate AI and machine learning researcher based in Chennai, India. She is pursuing a B.Tech in Computer Science and Engineering at SRM Institute of Science and Technology alongside a BS in Data Science and Applications at the Indian Institute of Technology Madras, and has worked as a research intern at IIT Madras. Her published research spans 3D scene reconstruction, EMG-based gesture recognition, and machine learning applied to scientific and healthcare problems.
Background
Originally from Andhra Pradesh, Keerthika has lived across several states in South India before settling in Chennai for her studies. Balancing two demanding degree programs at once, she gravitated early toward research — drawn to problems that sit at the intersection of multiple fields rather than fitting neatly into one domain. Her earlier research experience includes intent-based networking and computational chemistry, both explored during her time as a research intern at IIT Madras.
Research
Her first peer-reviewed paper, “A NeRF-Transformer Hybrid Framework for High-Quality 3D Scene Reconstruction and Contextual Interpretation”, was published at the 2025 IEEE International Conference on Smart and Intelligent Systems (SISCON 2025). The work combines Neural Radiance Fields with Vision Transformers to reconstruct photorealistic 3D scenes from sparse 2D images while adding semantic understanding of the objects within them.
Her second paper, “Cross-Subject Robust EMG-Based Hand Gesture Recognition Using a Hybrid CNN-Transformer”, has been accepted at the 17th International Conference on Recent Engineering and Technology (ICRET 2026). It proposes a multi-scale CNN and Transformer architecture with supervised contrastive learning that reaches 92.4% cross-subject accuracy across 36 subjects — a step toward EMG interfaces that generalize to users they were never trained on.
Alongside publications, her project work applies machine learning to early-stage lung cancer detection from medical imaging, tracing sources of PFAS contamination in environmental datasets, and fine-tuning large language models on Kubernetes documentation for intent-based network configuration.
Direction
Keerthika plans to pursue a PhD in computer vision, natural language processing, or machine learning for scientific applications — ideally somewhere those areas overlap. She has completed NPTEL certifications in machine learning, data science, networking, and databases along the way.
Beyond research
Outside of research she sings, draws, and has trained in Indian classical dance for years. She speaks five languages and is currently learning French. If you want to talk research, AI, music, or South Indian food, get in touch.