I enjoy standing at the intersection of business (to interact with business stakeholders and senior leadership team), technology and math ( build optimization/machine learning models). My career passion is to help make product and business operations smarter through innovative and practical analytics.
I want to take my passion for the Data Science field to the next level by working with the dynamic people in the industry. I would like to contribute to the society through my knowledge in this domain.
Present Projects ▪ Developing a people counter algorithm using Computer Vision (Python, OpenCV) and a tracking algorithm to count the people moving in and out of the retail stores. The project includes creating a dataset, training, and validation of the algorithm Past Projects ▪ Trained an object detector using ResNet50 model (Python, OpenCV) for the pilot project of self-checkout stores to detect the fraudulent scans. Also,developed a scan activity detection algorithm using pose estimation. The model achieved 80% accuracy in detecting the fraudulent scans ▪ Planned and developed the human activity detection algorithm (Python, OpenCV) using a human pose estimation algorithm to understand the customer activity while they are handling the objects in the retail stores. ▪ Involved in the research project related to activity recognition and behavioral analysis for identifying customer activity in the retail stores using computer vision. Communicated the future project plans and presented the findings to the executive stakeholders by evaluating the research papers in the field of behavior analysis (Gait Analysis) ▪ Created and maintained fully automated CI/CD pipelines for code deployment using CircleCI
Worked on data collection strategies for the future Machine Learning implementations Performed sentiment analysis on the chat messages using Natural Language Toolkit(NLTK) and Naive Bayes algorithm Performed basic customer segmentation using clustering algorithms to understand the users Worked on building a chatbot using Rasa NLU Designed dashboards using Tableau to solve business problems related to customer acquisition and communicated the results to the executive stakeholders. Improved the chat response on the platform by over 60 % by analyzing customer behavioral patterns Built a Chatbot using Google’s Dialogflow (formerly api.ai) platform for the mobile application
Collaborated with the Marketing team to improve the customer acquisition strategies using design thinking approach Worked as an Agent on the chat based platform to understand the customer behaviour and reported the findings to the executive stakeholders
♦Concept generation and design thinking on new product development
♦Analysis of stakeholder requirements
♦ Design and prototype building of a product based on Thermal Paper Technology for RICOH, Japan
2015-04 - 2015-08
Finding Donors for Charity (Supervised Learning):
The goal of the project was to construct a model that accurately predicts whether an individual makes more than $50,000. (Random Forest, SVM and Naive Bayes Classifier) Tuned and improved the performance using Random Forest algorithm that resulted in a final accuracy of around 86% and the F1 score of 73%.