• Utilized a Generative AI with Google's Gemini model to accurately translate non-English responses in employee surveys, preserving contextual meaning and enabling the analysis of 100% of feedback across global markets.

• Leveraged few-shot learning prompt engineering with PaLM 2 Text Bison to accurately summarize text responses across multiple segments.• Analyzed Accenture's largest survey, TGPS, ensuring on-time delivery of dashboards and ad hoc requests from leaders.• Built a sentiment analysis tool using the Google Cloud Natural Language API in GCP and custom Python code to analyze employee sentiments across various segments and surveys, achieving 97% accuracy. Deployed the results in a… Show more • Leveraged few-shot learning prompt engineering with PaLM 2 Text Bison to accurately summarize text responses across multiple segments.• Analyzed Accenture's largest survey, TGPS, ensuring on-time delivery of dashboards and ad hoc requests from leaders.• Built a sentiment analysis tool using the Google Cloud Natural Language API in GCP and custom Python code to analyze employee sentiments across various segments and surveys, achieving 97% accuracy. Deployed the results in a Qlik dashboard for global clients.• Developed and Improved the custom topic modelling algorithm by using advanced coding techniques and generalizing the code, resulting in a significant increase in the efficiency and accuracy of the process.• Implemented GCP Language Detection API model to correctly predict the language of the text to replace the old method which significantly increased accuracy from 75% to 98%.• Worked on multiple ad hoc requests, including deep-dive analysis on employee Gallup data, multi-segment text and numeric analysis, triangulated analysis, work-life balance analysis, and more. Delivered actionable insights and suggestions for clients around the globe.• Developed and automated a methodology for representative score calculation and country bias removal using Python, significantly reducing processing time and ensuring timely delivery of client requests.• Led a team of 2+ data scientists on multiple projects and ad-hoc assignments. Show less

• Developed Anomaly detection algorithm in python Prophet including suggestions oncorrection factor to be applied to correct sales value and volume trend• Worked on Store Universe Estimation for African market• Worked on Data Validation for Retail Survey & delivered in depth insights on current trends.• Worked On Category-Brand- Manufacture Analysis for client in Canada Market

• Developed Market Segmentation to identify key characteristics of markets using advanced preprocessing of data and K- prototype clustering algorithm which helped strategic implementations to save cost and improve the product• Improvised the existing product and bring more idea on the current panel size by developing a tool in python which brought down execution time by 90% and helped further to improvise the cost of the product And delivered sizing requested by clients in the USA… Show more • Developed Market Segmentation to identify key characteristics of markets using advanced preprocessing of data and K- prototype clustering algorithm which helped strategic implementations to save cost and improve the product• Improvised the existing product and bring more idea on the current panel size by developing a tool in python which brought down execution time by 90% and helped further to improvise the cost of the product And delivered sizing requested by clients in the USA market• Developed the classifier model to replace the classic approach of the store classification stratification which brought more than 20% improvement on the CVs of RMS sample design• Developed pathway to run Standard Error calculation studies for clients and regular studies by automating end to end tasks in python for the Canadian market• Deployed RMS Sample Design on USA market and delivered it for Food, Drug, Convenience and Pet channels which helped in getting ideal sample size and reduce overall cost on the current panel of stores• Deployed existing products on the new market from scratch which brought more ad-hoc to the team• Delivered on time ad-hoc analysis to stakeholders with tight timelines• Arranged weekly meetings with stakeholders, Assured continued progress Show less