Hani Kancharla

Developed enterprise-grade conversational AI using GPT-4o and LLaMA-2 via LangChain, deployed via FastAPI on Azure AKS. Built Retrieval-Augmented Generation (RAG) pipelines with FAISS for context-aware document intelligence. Engineered LangChain-based autonomous agents for task automation, reasoning, and dynamic knowledge retrieval. Automated CI/CD pipelines for AI models using GitHub Actions, improving deployment velocity. Created interactive Streamlit dashboards for real-time LLM performance tracking and user interaction analytics.

Designed and implemented scalable pipelines for training and deploying large language models (LLMs), enhancing platform intelligence and user interaction capabilities. Developed reusable CI/CD workflows using GitHub Actions and MLflow for automated model versioning and promotion across development, staging, and production environments. Built high-performance model serving infrastructure using vLLM and Kubernetes, achieving millisecond-level latency for real-time inference. Partnered with data science teams to operationalize AI agents and RAG systems, leveraging vector databases and LangChain for advanced information retrieval, applicable to educational content discovery. Established model governance and approval workflows, ensuring security, compliance, and auditability for production AI systems.

Designed and developed deep learning models for automating complex logistics and security threat analysis. Engineered and optimized neural networks for real-time inference, focusing on low-latency requirements. Collaborated with cross-functional teams to integrate AI solutions into existing platforms, ensuring scalability and performance.