Modhasreya Lade

I've developed Explainable AI algorithms and edge-deployable intelligent systems at the Centre for AI Research and Autonomous Cyber-Physical Systems at Universitetet i Agder, spanning computer vision, embedded AI, and distributed learning.

I proposed Meta-Tsetlin Machine architectures for interpretable knowledge transfer and hierarchical learning, building reusable frameworks for training, benchmarking, visualization, deployment, and explainable machine learning pipelines. My CS-Meta-TM framework achieved up to 98.1% classification accuracy.

At the Centre for Human-Centric AI at IIT Madras Pravartak Foundation, I architected an LLM-powered ETL pipeline for 500+ government audit reports, reducing manual effort by nearly 80%. I also built PDF extraction, RAG, fine-tuning, automated Excel reporting, and conversational question-answering workflows.

My work includes a 13.5 MB lightweight YOLO thermal object-detection model deployed on Raspberry Pi, Rock Pi, and NVIDIA Jetson AGX Xavier, published in the IEEE Sensors Journal. I'm completing a B.Tech in Artificial Intelligence and Data Science at IIITDM Kurnool.