As part of the strategic Data and AI team in Government and Public Sector Consulting, Avishek is responsible for a variety of activities in the conceptualization, development, and expansion of solutions and technologies for EY. He led multiple initiatives through client interactions, development, and delivery. His work ranged across strategic initiatives in Advanced Analytics, Machine Learning, Artificial Intelligence and Big Data. He led initiatives to progress, create, deliver, and deploy AI based solutions across the Government landscape. This included collaborating with government agencies to understand their unique challenges and requirements, identifying opportunities where AI and data-driven technologies can provide value, and developing tailored solutions to address these needs.Currently working on a Central Government Cybersecurity project which includes log analysis to detect suspicious activities using advanced AI-ML based solutions and using the Elastic-Logstash-Kibina tech stack.

- Target Campaigning for a Leading EdTech Company: Data aggregation from multiple sources programmatically through API, Market Mix modeling, and dashboard for the client for easy data visualization. Google Cloud Platform native services.- Aadhaar masking solution for a Leading Bank - Use of AWS native services as well as non-native approach using Tensorflow, OpenCV to build two different approaches to compare cost optimization and accuracy trade-off.- Chatbot for a Banking client - Using OpenAI services.

Project Summary- Credit Card Default Prediction."Financial threats are displaying a trend about the credit risk of commercial banks as theincredible improvement in the financial industry has arisen. In this way, one of thebiggest threats faces by commercial banks is the risk prediction of credit clients. Thegoal is to predict the probability of credit default based on credit card owner'scharacteristics and payment history."This is a Machine Learning Project where a classification model needs to be built to best classify which of the Credit card owners are likely to default on their payment.This is an ongoing project.

This was a 3-month training and internship program hosted by IIT-Kanpur on Applied Machine Learning and Data Science by Professor Laxmidhar Behera. The program was divided into 9 modules starting with Hands-on with Python, covering Machine Learning and Deep Learning, and ending with a Capstone project where we were divided into several groups.PROJECT SUMMARY-We were a group of 5 people and we chose the Zomato Restaurant rating as our project. Our goal was to predict the price of the restaurant by taking a few parameters as inputs and giving the output. Initially, the data was not clean so we had to do some feature engineering and feature selection, and data scaling as part of EDA. Then we tried many Machine Learning Models to get the model with the highest accuracy. Also, we tried hyper-tuning the models. This was a regression problem statement so we chose RMSE and MAPE as our evaluation metrics. Finally, we chose the Adaptive Gradient Boosting model as it gave the highest accuracy of 91%.Then the model was deployed in the Heroku cloud platform by using the Falsk framework and HTML and CSS for the frontend designing.The model is up and running and the entire project can be seen in the Github with the following link: https://github.com/avishekdeb13/zomato-rating-prediction