Harsita Mav
@harsitamav
Machine Learning Engineer at Metamorphosys Technologies
Mumbai, India
Harsita Mav is a Machine Learning Engineer with a background in developing AutoML frameworks, fraud analytics, and high-accuracy OCR systems. She has experience leading data science teams and conducting research in genetic algorithms for optimization. Harsita holds a Bachelor of Technology in Information Technology and has demonstrated technical proficiency in Python, deep learning, and product strategy through various industry projects and academic research.
Experience
Machine Learning Engineer
Metamorphosys Technologies
Established and led the Data-Science team at Organization, developing a Configurable AutoML Framework that slashed model development efforts by 80%. Engineered backend ML models for fraud claims adjudication achieving an 85% accuracy rate. Delivered a Fraud Analytics feature for a Hong Kong client within a 1-month deadline. Developed an OCR system with 98% accuracy for text and table extraction. Orchestrated code standardization for GenAI, OCR, and Risk Summary modules. Leveraged Django, Python, Postman, and MongoDB to detect fraudulent claims. Spearheaded integration of Generative AI and LLMs for automated report summaries.
Research Assistant
K.J. Somaiya Institute of Engineering and Information Technology
Developed ML based Genetic Algorithm for Optimization and generation of rules on 5 types of datasets. Optimization and Visualization of the results were done using Scikit and Matplotlib libraries. Conducted research under the mentorship of Dr. Vijaya Pinjarkar.
Software Developer Intern
Digel Systems
Conceived and crafted a comprehensive web application tailored for efficient Pizza Parlour Management. Architected and implemented critical features such as a secure Login module, Inventory Management system, and integrated Bill generation module using Python module Tkinter.
Education
K.J. Somaiya Institute of Engineering and Information Technology
Bachelors of Technology
Information Technology