Vaibhav Kesri

Highly motivated individual committed to hard work and intensive analysis. Flexible and polished AI/ML Data Scientist promoting well-developed skills in machine learning, deep learning, natural language processing, Knowledge graphs and Big Data.

• Knowledge Graph from Informal Text: Architecture, Components, Algorithms and Applications, Chapter in book : Application of machine learning,2020.
• Knowledge Graph based Automated Generation of Test Cases in Software Engineering, CoDS COMAD 2020: 7th ACM IKDD CoDS and 25th COMAD.
• AutoKG- An Automotive Domain Knowledge Graph for Software Testing: A position paper, IEEE International Conference on software Testing, ICST 2021.
• Probing the SpanBERT Architecture to interpret Domain Adaptation Challenges for Coreference Resolution, CEUR Workshop Proceeding ,AAAI 2021 February 9, 2021.
• An Unique Solution for N queen Problem, International Journal of Computer Application 43(12):1-6, April 2012.

Deep Learning Specialisation, offered by Deeplearing.ai

• Designed and implemented transaction-specific features to enhance the performance of the Smart Authentication model:

Created graph-based embeddings capturing neighborhood information to enrich domain-specific entities. 

    Current Status: Scheduled for integration into the final model for the APMEA region.

    Developed velocity features utilizing browser information from 3DS transactions to identify anomalies in transactions.

    Current Status: Scheduled for integration into the final model for the APMEA region.

    Created seasonality-based features to improve the downstream model's predictive performance.

    Current Status: Under testing to evaluate feature efficacy.

     Role: Senior Data Scientist, Organization : Mastercard

• Designed and implemented a predictive model to classify online transactions (3DS 2) into frictionless or challenge routes:

    Improved transaction flow by reducing unnecessary challenges while ensuring security and compliance with 3DS 2 protocols.   

    Generated probability scores to enhance decision-making and optimize customer experience.

    Technology Used: Hadoop, Pyspark, Python.   

     Role: Senior Data Scientist, Organization : Mastercard

• Identify customer cohorts that have intent to buy various plans in Airtel Black: To create a sample from the population, for exploratory data analysis, domain-specific feature engineering, and feature selection followed by multi-class Model creation for imbalance data, optimization, and evaluation of the model.

Technology Used: Hadoop, Pyspark, Python, Organization : Airtel

Role: Data Scientist, Organization : Airtel

• Graph Data Base creation for different down streaming tasks: To fetch user-related data from multiple sources and create a Knowledge Graph for multiple downstream tasks for example product recommendations and churn prediction etc.

Technology Used: Hadoop, Python, Neo4j, Cypher, Organization : Airtel

Role: Data Scientist, Organization : Airtel

• Manage prepaid to postpaid customer acquisition pipeline: To manage Hive tables and pipeline to target potential prepaid customers for postpaid acquisition.

Technology Used: Hadoop, Pyspark, Python

Role: Data Scientist, Organization : Airtel

• Knowledge Graph creation from structured, semi-structured, and unstructured data: The Target is to help testers in executing testing by giving them insights into various steps happening behind the functionality and help to provide the correct sequence of steps to test a feature. It is a mix of NLP, predictive models, domain ontology creation, and knowledge graphs.

Technology Used: Python, Numpy, LSTM, Keras, Sklearn, pandas, neo4j, Stanford core NLP, spacy, etc

Role: Researcher (Senior Data Scientist), Organization : Robert Bosch (RBEI)