Kuldeep Soni

Experience

• 

  Syntel

  May 2016 - Jun 2016

  Data Science Intern

  text miningproject -1:- comparision of two facebook pagesextract the data of two fb pages amex and visa pages and carried out the the analysis based on that text. used the topic modeling ,trend analysis ,wordcloud technique to extract the useful result . compare the pages based on sentiment analysis and fitted the best model among naive base , random forest.project 2: keyword prediction from text data That is the kaggle project. where naive base , random forest and support vector machine techniques used to fit the data Show less

• 

  GEP Worldwide

  Jun 2017 - May 2019

  Data Scientist

  1. Automation of Spend Category Classification (text data) of Operations: • Built a classification engine for pre-existing clients • Advanced text cleaning techniques such as NER using Python3 • Machine learning techniques for classification used such as Support Vector Machine (SVM), Naïve Bayes, Logistic Regression and Random Forest • Optimized Elastic Search Engine queries to search from historical data • Developing solutions for cold-start clients2. Supplier De-Duplication Engine: • Developed an engine to provide normalized names of each supplier from a collection of inconsistent supplier names, to ease report generation • Used different types of distances such as Jaccard, Levenshtein and Sequence matching and NLP techniques to clean the data.3. GEP Guided Buying: • Predicting the required category of product/service for the clients, based on their input queries • Developing supplier and catalog ranking algorithms based on the queries and predicted categories • Used Topic modeling, Natural Entity Recognition (NER) and ML models to build the category predicting algorithm4. Introduced a chat-bot concept in GEP Reporting Tool: • Developed an initial prototype chat-bot for handling client queries relating to their spend data and provided Proof of Concept. • Used advanced Natural Language Processing (NLP) concepts to translate queries written in plain English into required reports5. Extracting summary information from long PDF/Word contracts: • Chosen project for internal competition, Hackathon 2018; now an official project. • Extracting start/end dates of contracts, contract parties, clauses, renewal terms, etc. • Completed within time limit of 24hrs. Won 2nd place for submission. Show less

• 

  Dell Technologies

  May 2019 - now

  • Senior Advisor, Data Science

    Mar 2024 - now

  • Advisor, Data Science

    Apr 2021 - Mar 2024

  • Senior Analyst, Data Science

    May 2019 - Apr 2021