Driving innovative solutions across domains by leveraging deep learning, machine learning, and Generative AI to tackle complex business challenges

Organized the 2024 Radboud Career Event as a Board Member of the BBB Foundation.

Master’s Thesis - Leveraging LLMs for generating Local Weather Reports– Developed a comprehensive pipeline using LLM and RAG for generating location-specific weather reports. Integrated a pilot chat application to facilitate user interaction with the generated reports.– Implemented ELT processes to effectively gather and process data from 5 weather stations across the Netherlands.– Evaluated with experts achieving high scores, demonstrating strong location-specific accuracy and positive user acceptance.Research Internship - Error Analysis of Whisper-large-v2 based Automatic Subtitling Pipeline– Analyzed manual and AI-generated subtitle data to optimize speech recognition, punctuation restoration, and subtitle segmentation within the whisper-based pipeline.– Employed quantitative and qualitative methods, including temporal alignment of whisper-generated and manual subtitles, and estimation of confidence using average log probability to assess whisper uncertainty.– Identified challenges such as error in punctuation prediction, and substitution errors, highlighting system limitations across diverse linguistic contexts and indicating balanced performance with an EER of 0.29.