HariChandhana Kalikiri

Developed POC for implementing RAG for Financial Forecasting and Business Analyst Bot.

1. ImperialDade :

• Led an NLP project utilizing LLM models like DeBERTa, using Databricks, PySpark, and Pandas.
• Managed products and vendors from 49 acquired companies, clustering similar products and creating golden records.Leveraged record linkage and clustering techniques, while also creating features such as color and material using DeBERTa and RoBERTa LLM models.
• We utilized GENAI through LangChain for precise dimension extraction from product text, enhancing data accuracy to enhance clustering accuracy and remove ambiguity.

2. ARS :

• Built a sophisticated demand forecasting model using c forecasting accelerator, aimed at predicting booked, bookable, and leadsrun demand. The model leverages advanced time series analysis and statistical modeling techniques.
• Achieved accurate demand trend predictions, significantly supporting strategic decision-making and optimizing resource allocation..

3. T-mobile :

• Developed an architecture to determine which network KPIs influence customer activations and deactivations at the CBG level, utilizing weighted customer mobility data for causal analysis.
• Collected comprehensive network performance metrics (coverage, capacity, call drops, throughput, latency, etc.) and customer metrics (activations, deactivations), including mobility data.
• Applied customer mobility data to weight activations and deactivations, accounting for the influence of mobility patterns on these metrics.
• Employed panel and fixed effects models to analyze the causal impact of network KPIs on activations and deactivations, refining insights into the effects of network changes.

• Utilized product data to develop and implement a sales forecasting model,Analyzed historical sales data, product features, and market trends to forecast future sales accurately.
• Worked on dealer app,workflow involves warranty of products, claims, coverageMongoDB, Vuejs, Nodejs, were used to develop the application and AWS Amplify for deployment.

• Designed and developed a proof of concept (POC) for a plagiarism detection model.
• Created a classification model for sentiment analysis of product reviews, employing machine learning algorithms to analyze and interpret customer feedback, providing actionable insights into product performance and customer satisfaction.
• Developed a classroom application using VueJS, NuxtJS, and Tailwind CSS to create a responsive and user-friendly interface.
• Implemented a Query Resolution System that automatically maps and routes queries to appropriate employees, optimizing query handling and response efficiency.

• Submitted POC on Detecting Performance Anomalies in Cellular Networks and developed an alert generating application.
• Developed KPI's and presented it to management which led to execution plan.