Veeranna Kondapalli

Architected a Retrieval-Augmented Generation (RAG) system: To synthesize years of unstructured sales notes, store-opening agreements, and monthly visit reports into high-level account summaries.

• Engineered an automated "Next Actionable Insights" engine using LLMs enabling sales representatives to instantly grasp historical context and pain points before account visits.
• Developed a scalable dual-service architecture utilizing FastAPI for high-performance backend logic, and Streamlit for an intuitive, interactive user interface.
• Implemented containerized deployment using Docker, ensuring environment consistency across development and production, and facilitating seamless transitions during personnel turnover.
• Optimized information retrieval by building a searchable knowledge base with a vector database, reducing sales representatives' preparation time by providing contextually relevant historical data.
• Tools Used: Python, FastAPI, Streamlit, Docker, LangChain, Vector Databases, OpenAI API, and NLP.

Competitive intelligence:

• Engineered a competitive intelligence framework to identify and monitor Eastman film brands across diverse global marketplaces.
• Analyzed competitor and customer behaviors to generate actionable insights, enabling data-driven strategic decisions that increased revenue and expanded the customer base.
• Performed regional market gap analysis (oversaturation vs. Undersaturation, identifying high-potential areas for business expansion, and resource allocation.
• Tools Used: Python (Scikit-learn), SQL, Power BI, and Machine Learning Models.

Customer Segmentation:

• Segmented the LLumar and SunTek customer bases using K-means clustering to drive targeted marketing and sales strategies.
• Developed a robust RFM (Recency, Frequency, Monetary) data architecture based on two years of historical transactional data.
• Optimized data quality by addressing skewness and performing normalization to ensure high-accuracy clustering results.
• Tools & Methods: Python (Pandas, NumPy, SciPy, Scikit-learn), MySQL, Power BI, Statistical Hypothesis Testing.

• Architected and deployed scalable web scrapers to extract high-fidelity vehicle specifications (model, trim, color), and multi-angle imagery from Tier-1 manufacturer sites, including BMW, Ford, and Porsche.
• Tools used: Python (BeautifulSoup/Selenium).
• Automated complex ETL workflows by migrating manual data tasks to Python scripts, streamlining data ingestion for DIY Gila retailers into central databases.
• Developed real-time Tableau dashboards to visualize weekly sales performance and inventory levels, reducing manual reporting time, and improving stock-level accuracy.
• Tools Used: Python, SQL, Power BI.