Jatin Kayasth

Machine Learning Researsher Intern

S&P Global•  June 2023 - August 2023

• Reduced processing time by 80% and increased data accuracy of deep learning models by 10% by developing a data preprocessing pipeline for collecting and cleaning precipitation data from NASA • Achieved an accuracy rate of 65% by researching and implementing cutting-edge machine learning algorithms to develop SRGAN models for transforming low-resolution images into high-resolution images

Data Scientist

Infosys•  December 2018 - June 2022

• Improved Company’s rank on SuperGLUE tasks from 15th position to 8th position by fine-tuning models such as GPT2, GPTNeo, BERT, RoBERTa, ALBERT for text summarization, text generation, question-answering systems • Reduced manual data entry task by 70% while developing and managing the entire backend for applications such as Text analytics program using HMM and Viterbi Algorithm, Named Entity Recognition for resume • Achieved f1-score of .93 to detect credit cards fraudulent transactions by building a machine learning model capable of handling huge class imbalance • Collaborated with cross-functional teams to integrate real-time monitoring tools with predictive alerting capabilities, resulting in a 30% increase in proactive issue resolution