Kotapati Lakshmareddy

Generative AI Engineer and Data Scientist with 2 years of experience building LLM-powered applications, RAG systems, AI chatbots, and enterprise AI solutions. Skilled in Python, Azure OpenAI, LangChain, LlamaIndex, FastAPI, SQL, and Vector Databases. Experienced in developing production-grade AI applications, AI agents, document intelligence solutions, and automation systems across banking, manufacturing, and ERP domains.

Generative AI Engineer | Data Scientist - QX Impact

(2025-09)

• Engineered automated data validation framework for production environment, ensuring 100% data completeness and accuracy, reducing manual verification time by 90% and eliminating 95% of human errors through trigger-based ERP executions
• Developed an Azure OpenAI–powered ERP table mapping solution that automates source-to-target mapping recommendations from metadata, successfully identifying 50% of target table mappings and reducing manual mapping effort.
• Implemented fine-grained user and module-level database access control using SQL GRANT/IMPERSONATE permissions for 50+ users, improving data security and compliance
• Developed Retrieval-Augmented Generation (RAG) chatbot using LLMs, LangChain, and vector databases, improving knowledge retrieval accuracy by 75% and reducing manual information lookup time
• Optimized chatbot response quality and reduced latency by 60% through prompt engineering, chunking optimization, and embedding tuning
• Integrated Azure Key Vault, Azure Blob Storage, and SharePoint with FastAPI for secure, scalable data validation systems

Generative AI Engineer | Data Scientist - Care Analytics and Advisory Private Limited

(2024-08 - 2025-08)

• Deployed production-grade WhatsApp banking chatbot using Meta API and Rasa NLU, serving 1000+ active users with 95%+ query resolution rate and reducing customer support costs by 40%
• Developed automated news sentiment analysis using LLMs, processing 500+ financial articles daily with 92% accuracy, providing actionable market insights and reducing manual analysis time
• Engineered credit risk models (PD, LGD, EAD) on financial datasets, improving loan default prediction accuracy by 35% through comprehensive EDA and feature engineering using Pandas and NumPy
• Built AI-driven video summarization using Mistral LLM, automatically generating accurate summaries from 1000+ hours of content with 88% coherence score
• Built intelligent document Q&A systems using RAG pipelines (LangChain, FAISS, ChromaDB), enabling context-aware responses from 500+ documents with 89% relevance score
• Engineered Speech-to-Text (STT) pipelines using OpenAI Whisper and FastAPI to capture and transcribe user voice inputs
• Researched and applied advanced GenAI techniques, experimenting with multiple LLM architectures to optimize accuracy and efficiency

B.Tech - Computer Science And Engineering - IIIT BHAGALPUR (2020-09 - 2024-04)