Jyothi Krishnakumar

Utilized optimization techniques to determine machine cut quantities, optimizing reel size based on demand, and developed a paper requirement algorithm considering factors like gsm, streamlining procurement and production planning processes for cost-efficient operations in a supply chain optimization project. Analysed and modelled 4+ years of monthly data for 500+ SKUs, selecting the best-performing model based on model performance metric for demand and stock planning. Developed a dynamic price quotation system utilizing Power Apps and Power Automate, enabling user-defined carton and location specifications. Integrated a hierarchical approval workflow, resulting in reduced lead times and enhanced sales team efficiency. Performed ETL transformation leveraging python scripting and API mechanisms, to develop multiple Dashboards in Power BI and report KPIs to leaders for efficient decision-making. Performed YOLO object detection on custom data to develop an image identification solution for validating daily image submissions by sales representatives, ensuring data accuracy and integrity. Automated cost variability analysis of P&L tables using Python, enhancing operational efficiency. Integrated ChatGPT with Power BI to generate concise summaries of P&L cost fluctuations, facilitating data-driven decision-making.

Engineered a NLP Query Engine tool leveraging NLP and Python to autonomously derive data insights based on user queries, enhancing data-driven decision-making. Developed a Plotly Dash-based web interface for user-friendly waybill entry and updates, integrated with Azure Microsoft SQL for robust data handling, fostering streamlined operations and data accuracy. Executed efficient and secure data migration and integration protocols employing REST API, AWS S3, MongoDB ensuring seamless data flow and storage management. Automated manual target management processes within sales and operations tool, leading to substantial time and resource savings.

Utilized Python to develop a web crawler that extracts information from Wikipedia info boxes to create databases for knowledge graphs. Built NLP-based Question-Answering system using the BERT model to deliver precise and contextual answers.

Leveraged SPSS software to execute meticulous data cleansing and management for a comprehensive Tuberculosis prevalence survey study. Conduct periodic data analysis using R programming to find insights and generate detailed reports for catastrophic expenditure studies of patients to provide data-driven suggestions for policy implementation.