Anindita Deb

Buffalo, New York • aduzuz@r.postjobfree.com • 716-***-**** • LinkedIn• Github

EDUCATION

State University of New York at Buffalo - Buffalo, NY Aug 2021 - Feb 2023 MS in Computer Science and Engineering GPA: 3.4

Relevant coursework: Statistical Data Mining(Supervised and Unsupervised), Deep Learning, Analysis of Algorithms Veer Surendra Sai University of Technology - Odisha, IND Aug 2011 - July 2015 Bachelor’s in Electronics and Telecommunication GPA: 3.71 EXPERIENCE

Virtusa - Pune, IND Aug 2020 - May 2021

Senior Consultant

•Conducted a data regression analysis of the relationship between company stock prices and industry trends achieving a 15% more accurate prediction of performance than in previous years.

•Created SQL and Python programming modules and Tableau reports for custom insights required by our product team.

•Cleansed and manipulated datasets using EDA to gather insights, patterns, and present findings in an easy-to-complex format. Accenture - Bombay, IND Aug 2015 - July 2020

Data Analyst

•Developed Stored Procedures to implement complex business logic related to computing Insurance Policy Rules.

•Developed meaningful data insights using data management tools such as Informatica and visualization tools such as SAP Business Objects required for strategic banking decisions such as loan approvals.

•Designed automation scripts to fetch data from the input streams, reducing the overall execution of downstream jobs from 1 hour to 30 minutes.

•Assisted the business and technical staff with the monthly development, testing, and maintenance of the 2+ new programs.

•Contributed to deployment activities significantly, using CI/CD tools GIT, Bitbucket, and Jenkins.

•Developed Fast load, and TPT scripts in Teradata to automate data extraction process.

•Cross-trained about 50 new team members which enhanced their productivity within a few months by 20 % and ensured seamless service.

•Facilitated team-building activities to ensure effective communication among team members. PROJECTS

Damaged Hair Cell Count Detection May 2022 – Nov 2022

• Developed an automated system in PyTorch to capture damaged cell regions spread across ring-like structures in hair follicles which has the potential to aid in the correct medical diagnosis of possible tumors.

•Cell regions were captured by template matching, overlapping bounding boxes were reduced using NMS suppression.

•The results of NMS suppression were further fed to DBSCAN algorithms that resulted in the segregation of cells from noncells.

•Trained a Resnet Architecture on the input (from the above step)and a context related to every input, which is a larger area of interest containing neighboring cells of the input to identify the cells more accurately.

•Implemented hard mining that aids in identifying the wrongly classified samples from one iteration and weighing them more while randomly sampling data in successive iterations of the training period.

•Achieved an overall accuracy of 91 % and recall of 97 % on test data. Developed Credit Risk Scores Aug 2022 - Nov 2022

• Performed critical data mining using existing statistical tests such as Chi-square, and Anova -F to understand the relation of customer attributes to their payment behavior.

•Computed statistical significance of each independent variable such as loan status, home ownership etc, using existing statistical metrics such as weight of evidence and information value.

•Implemented a probability at default model using Logistic Regression using the Scikit - Learn library and evaluated the model performance based on the Gini score and Kolmogorov – Smirnov curves which were 0.6 and 0.4 respectively.

•Further the parameters/coefficients of each feature from the above model were used to compute the credit risk scores.

•Experimented with Random forests(96 percent) on the same data and it outperforms Logistic Regression(94 percent) on test data by 2 percent however Logistic Regression model being simpler in terms of interpretability further aids in evaluating customer features such as occupation, property, etc for strategic banking decisions such as loan approvals. SKILLS

Languages: Python, SQL, R, Javascript

Tools: Tableau,SAP Business Objects, Matplotlib, Seaborn, Git, Bitbucket, Jenkins, Informatica, AWS, Teradata Frameworks: Pandas, PyTorch, Scikit - Learn, Flask, Docker

Contact this candidate