Harikrishna Dev

About

As a highly motivated and results-driven Business Analytics professional, I specialize in using data-driven insights to help companies make critical decisions and drive business growth. With a Master's degree in Business Analytics from The University of Texas at Dallas and a Bachelor's degree in Mechanical Engineering from National Institute of Technology Karnataka, I bring a unique combination of analytical skills and technical expertise to the table.

During my tenure in the Analytics industry, I have worked with different companies and looked at problem statements from a Business and statistical point of view. I worked with Sam's Club and Walmart Customer Science teams to analyze customer behavior, and design data pipelines and tables. I worked in Flipkart's Customer Experience team, where I analyzed different Incident Management Systems and their impact on Customer Satisfaction Indexes.

Course work

Academic courses

• Advanced Statistics for Data Science
• Database Applications for Business Analytics
• Business Analytics With R
• Predictive Analytics for Data Science
• Programming for Data Science
• Applied Econometrics, and Time Series Analysis
• Operation Research
• Applied Computational Methods in Mechanical Sciences
• Numerical Methods in Engineering
• Applied Machine Learning
• Applied Deep Learning
• Business Data Warehousing
• Prescriptive analytics
• Applied Natural Language Processing

Online courses

• R Programming A-Z™: R For Data Science With Real Exercises!
• SQL - MySQL for Data Analytics and Business Intelligence
• Hadoop in Real World
• IBM: Tools in Data Science
• IBM: Data Science Methodology
• Programming for Everybody (Getting Started with Python)
• Python Data Structures

Experience

Daifuku North America Holdings Corporation

IT Analyst Intern

• Developed proprietary Generative AI tool to perform text mining on report documents and provide source citations by leveraging the HuggingFace API and Natural Language Processing, reducing 2 hours of daily work
• Augmented the Paint Line assembly efficiency by 48% by identifying bottlenecks and forecasting recurring issues in the assembly line using Regularized Linear modelling techniques such as Elastic Net and Facebook Prophet
• Designed a Power BI dashboard to visualize and track employee performance metrics, which saved senior leadership 3 days of work every month by connecting data sources to Microsoft Azure

Flipkart Internet Pvt Ltd

Business Analyst - Customer Experience team on Agent Serve Charter

• Conducted comprehensive analysis of customer experience key metrics trends, providing senior leadership and product managers with actionable insights to drive product development, process enhancements, infrastructure requirements, root cause analysis, and integrated marketing and monetization strategies
• Deployed Customer Escalation prediction framework to prevent social media escalations based on user journey and pain points using Machine Learning techniques such as XGBoost and CV, which resulted in 27% decrease in escalations and 8% enhancement in Customer Satisfaction and Net Promoter Scores
• Collaborated on A/B testing initiative on e-commerce sales page during the `Big Billion Day` event, resulting in 10% increase in customer acquisition and 7% boost in average order value, driving ₹3M in additional revenue

Tredence Analytics Pvt. Ltd.

Senior Business Analyst - Sams Club Customer experience team

• Analyzed and reported on Overall Customer Experience on the e- commerce app and website using derived metrics consumed by senior leadership regularly to make product changes.
• Orchestrated data-driven discussions with Senior leadership to evaluate Experience Metrics; guided strategic decision-making for delivery modes and Supply chain systems, leading to a cost savings of USD 20,000 per year for Sam's Club.
• Worked with Senior leadership on major changes to the Experience Metrics. This helped the company make critical decisions on different modes of delivery, and changes to the Supply chain systems.

Business Analyst - Walmart Customer Science - Data engineering

• Developed and implemented multiple high-volume data pipelines and tables in terabytes to support data-driven decision-making, enabling timely and accurate assessment of customer behavioral patterns and identifying strategic growth opportunities.
• Developed automated notification pipelines to ensure smooth operation of data warehouses.

Business Analyst - Walmart Customer Science team

• Assessed customer behavioural patterns based on their buying capacity, retention period, and effect of Online Grocery. Built multiple data pipelines and tables to support related decisions.
• Determined the accuracy of sources of credit card transactional data and designed an internal mapping algorithm to understand customer spend behavior. The project’s objective was to identify potential customers to acquire into the Walmart environment using behavioral patterns and “Share of Wallet” data.

Personal Projects

Auto Regression Machine Learning Web application

This projects was to understand how to deploy Machine learning Web applications with Streamlit

• The website is hosted on Streamlit Cloud servers and has a simple python backend
• The AutoML functionailty works on PyCaret and outputs the best Regression model for the given data.
• Tools: Python, Streamlit, HTML

Store Sales Forecasting using Time Series methods

University of Texas at Dallas

This project examines the impact of gas prices on the economy and daily wages of the population in Ecuador, a country known for its reliable oil resources.

• Our regression model suggests that oil prices have a significant effect on sales.
• By including these variables in our model, we can better understand the relationship between these factors and sales, leading to more accurate predictions and better decision-making.
• Therefore, incorporating additional factors into our model and exploring alternative modeling techniques can help us obtain more accurate and reliable predictions.
• Tools: Python, STATA

Prediction of Heart disease using Classification models

University of Texas at Dallas

This project predicts if a patient will be diagnosed with heart disease based on health records.

• Our objective is to identify high-risk patients for CVD based on their current health-related information obtained from yearly checkups.
• The prediction will assist medical professionals and self-assessment tools to evaluate patients' course of medication and health plan choices.
• After running multiple binomial classification models, we were able to conclude XGBoost Classifier was the best fit in our case
• Tools: Python, H2O ML Platform

eCommerce RFP Validation for Walmart

Tredence analytics Pvt Ltd

Determining the accuracy of multiple sources of credit card transactional data to evaluate Market Share and Share of wallet of different eCommerce platforms

• The project consisted of creating a mapping algorithm to evaluate transactional accuracy of sample group.
• It also included identifying customers who can be attracted into the Walmart environment using behavioural patterns and Wallet share distributions of omnichannel customers.
• This helped us in determining the platforms strength in different shopping categories.
• Tools: R, Python, GCP BigQuery, SQL

Investigation of effect of different cutouts on composite laminates

Under Dr. S M Murigendrappa

Developing a mathematical model in the virtual environment and validate the mechanical properties of a composite beam with cutouts

• My major project thesis opened my understanding of the world of Composites, which showed me the use of uni-directional and bi-directional properties in different applications.
• A strong correlation between the tensile and vibrational properties of a laminate and fiber and notch characteristics can be observed.
• My team and I started our project by learning the different methods of fabrication of Epoxy Resin Bi-Woven Glass fiber composites.
• This helped me understand the fabrication process in composites like Carbon fibre and Glass fibre and the precautions taken to avoid fiber damage during processes like drilling.
• My proficiency in LabVIEW helped in the vibrational experimentation of the laminates. We were able to obtain simulation results that were very close to the experimental results with minimum error.
• We were also able to predict the crack initiation points and stresses using Hashind Damage Criteria with a good amount of accuracy.
• Tools:MATLAB, ABAQUS, GNU Octave

Fabrication and analysis of suspension system

Under Dr. Veersheety G

Design and fabrication of a suspension for a mini BAJA vehicle

• Design and Analysis of the Engine mounts, which included the Testing of rubber pads and aluminum mounts, Failure testing of engine bolts, Vibration transfer analysis to the engine using different materials and designs, etc., Numerical modeling and Experimental testing of Stress analysis and Fatigue on the MIG-Welded joints on the chassis and suspension, DFMEA and PFMEA on the vehicle for the event, etc.
• Tools:MATLAB, MSC ADAMS, Lotus suspension system

Implementation of a small scale smart city

Under Prof. K V Gangadharan

Building a small scale model of a smart scale model of smart city using open source tools and portraying the benefits of unifying the databases

• The mini project afterward in the course was a miniature model of a Smart Self-sustainable city where we implemented everything taught in the course.
• We built three autonomous bots and 4 self-sustainable buildings using Arduinos as CPUs and solar panels as energy sources, which were connected to a Master Central communication Unit (Raspberry Pi).
• The CCU displayed the information pertaining to the city like the position of the bots, air conditions of the city, emergency distress signals from each building, and computation of the shortest path for the ambulance bot to the given building, etc.
• This would be hosted on a webpage that is accessible by the students and staff when connected to the college network.
• Tools:Arduino, Python, Rasberry pi

Skills

Contact

• LinkedIn

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';