Govindu Sai Anish

Multimodal AI Agent for Autonomous Computer Use

• Developing a vision-language agent for autonomous GUI interaction using multimodal reasoning, GUI understanding, grounding, and action prediction. • Designing GUI grounding and action prediction to map instructions to UI elements and executable actions with error recovery. • Fine-tuning pretrained vision-language models with GUI-specific data and evaluating task success.

QuizApp Microservices - Spring Boot Microservices Project

Developed a Java Spring Boot microservices-based Quiz Application with independent Quiz, Question, API Gateway, and Eureka Service Registry services. • Implemented REST APIs and inter-service communication using OpenFeign, with Eureka-based service discovery and API Gateway routing for distributed service communication. • Integrated MySQL with JPA/Hibernate using layered Controller, Service, Repository, and Model architecture.

CortexRAG – Knowledge Base AI Chatbot (RAG System)

• Designed and implemented a production-grade RAG system over PDF, DOCX, and TXT documents, improving answer relevance by 35% and reducing hallucinations through grounded responses. • Engineered a hybrid retrieval pipeline (FAISS + BM25) with cross-encoder reranking (MS MARCO MiniLM), increasing retrieval precision by 25% and reducing irrelevant context passed to the LLM by 40%. • Developed an end-to-end system with LLM-based query expansion and a FastAPI + React architecture, achieving sub-2s response latency and improving retrieval coverage by 20% for ambiguous user queries.