At Gytworkz Technologies Private Limited, I used YOLO and Faster R-CNN to detect and classify bacterial colonies in petri dish images.
I designed and deployed a system to count and classify colonies by size, shape, and color, reducing manual counting errors and increasing processing speed. I also worked on preprocessing and model optimization to improve accuracy and reliability.
For a mental health analysis project with Prof. Udit Satija at IIT Patna, I developed an Operational Neural Network architecture to classify audio recordings into three depression categories, achieving over 0.95 classification accuracy. I also developed an application for people to analyze and track their mental health.
I designed a sentiment analysis project using central architecture and contrastive learning to classify reviews and explain the reasons for the classification. I also worked on a pizza sales analysis project using SQL to examine sales performance, customer preferences, and business trends.
