Comorbid Analyzer for Diabetes

• Developed a machine learning model to detect diabetes and related diseases without the need of health reports.

• Utilized patient data and medical history for accurate predictions.

• Implemented feature selection and preprocessing to enhance model performance.

• Curated a unique ensemble architecture and achieved 95% accuracy overall.

• Presented the project in a AICTE backed project expo with 400+ participants, enlightening interest among participants towards machine learning and depicting the capabilities of AI in the medical domain. 

• Published a research paper for making this model accessible for usage and further improvement with International Journal of Research.

​

Smart Learning Advisor

• Created a system that provides personalized learning style suggestions based on student inputs.

• Analyzed data such as subject preference, previous scores, and time availability to recommend optimal learning methods.

• Utilized decision tree and random forest models for predictive analysis with 90% accuracy.

• Integrated the model with Flask API and React.js Interface for usability.

• Presented the model in Hack-A-Bot RPA hackathon, capturing the interest of mentors and peers.

• Deployed the model into UI Path Orchestrator for enterprise level usage, making it accessible for mentors in order to guide their students using honeypot API.

​

Focus Flow

• Working on an AI based web application which serves as deviation and addiction-free social media platform.

• Backed by NLP models, Video and Image processing technologies, React.js and Firebase.

• Aiming to avoid distractions and extensive non-progressive usage of social media and promoting children friendly content for minor users by 90%.

• Driving the application as a user-centered platform, by leveraging comments as additional keys for filtering deviating and harmful content.