Tanya Mishra

I’m an AI & ML professional specialising in computer vision and real-time video analytics systems. I focus on building vision-based AI solutions that operate reliably in production environments, with strong hands-on experience in object detection and video understanding.

Most recently, I worked as an AI&ML associate at Avkalan.ai (Remote), where I built and fine-tuned LLMs and generative AI models for real-world deployment. I contributed through data analysis, prompt engineering, and scalable AI solution delivery—staying practical about reliability and impact.

Earlier, I interned at Intel Corporation on AI-driven smart kitchen solutions for automated ingredient detection and recipe suggestions, integrating ML models to optimize meal workflows and enhance user experience. At the Ministry of Housing and Urban Affairs, I analysed urban data and applied ML models to generate insights for Smart City Varanasi, supporting improvements in urban planning, resource management, and citizen services.

My project work reflects my breadth across multimodal learning and graph-based reasoning. I built a multimodal deep learning framework for binary change detection using Vision Transformers and Graph Neural Networks, implemented a Temporal Graph Neural Network for autonomous driving event prediction with NuScenes data, and developed a Python-based cross camera player mapping system using object detection and feature-based matching.