2021 - Present
University at Buffalo
Relevant Courses: Machine Learning, Statistical Data Mining, Python Programming for Data Science, Bayesian Networks in R, Data Models Query Lang. Club Membership: Women\'s Empowerment and Leadership Club.
2015 - 2016
Center for Development of Advanced Computing
2012 - 2014
University of Pune
2022 - 2022
University at Buffalo
• Coordinated with 5 team members to apply sports analytics on NFL games and predicted punt return yardage to help Buffalo Bulls gain most yardage.
• Presented on AWS/NFL’s Big Data Bowl competitions held from year 2019 to 2022.
2019 - 2019
• Automated multiple banking processes on ABot (ABSA Bot) platform leveraging python and reduced clients’ manual work by 95% which resulted in approximately $30 million in sales in 2019.
• Coordinated with stakeholders to conduct requirement gathering, use case feasibility study and prepared process definition documents (PDD), and solution design documents (SDD) for 10+ processes to automate them 100%.
• Implemented analysis, design, coding, system testing, integration, and UAT in the RPA platform.
• Extracted text from PDF files deploying Optical Character Recognition for better performance and security in banking systems.
2018 - 2021
• Collaborated with a US-based banking company to provide Data Science expertise.
• Classified emails based on text for appreciation emails using CRF and Named entity recognition implementing Bidirectional LSTM-CRF.
• Facilitated a team of 6 in retaining a South Africa based banking client using automation and data science skills worth $2m.
• Served as an interviewer for Infosys and mentored 10 new team members via knowledge transfer sessions.
2016 - 2018
Cybage Software Pvt. Ltd.
• Produced and deployed client-based Interactive advertisement application utilizing object detection and recognition.
• Brainstormed and managed a team of 2 to train models on more than 10k images extracted from live video
advertisements using multiple techniques such as Dlib, OpenCV, TensorFlow and YOLO Darknet to detect customized objects and to make them user interactive.
• Researched and Optimized models using various techniques such as early stopping, stochastic gradient descent and different hyperparameter tuning improving accuracy by ~85% with a greater number of detections.
• Developed deep learning photo caption generator using Convolution neural network and LSTM to generate captions for unseen images.
2023 - Present
• Collaborated with the CEO and COO to link Amazon Associates data on user purchases with Mixpanel to enhance data processing, metrics, and visualization. The resulting visualization accurately depicted top users and products sold, which in turn helped to identify target audience and led to an 80% increase in sales.
• Developing a chatbot on the Craftle website by incorporating AgentGPT (autonomous AI agent), increasing average customer satisfaction rating by 8%.
Honors & awards
• Trained a Gradient Boosting Classifier model using weather API, distance API and traffic API data to develop an app to predict mode of transportation user should choose with an accuracy of 98% (15% improvement over various regression models). • Collected weather and traffic data using APIs by creating data pipelines and stored the data in CSV files and then combined the CSVs taking useful features and created labelled dataset. • Created impactful app which is being used by 3 million people every day and won Best AI/ML project title in Hackathon.