Buyyarapu Vishal

FastAPI, Google Gemini LLM, MCP Server, HTML/CSS/JS

• •Developed a modular, web-based banking chatbot using FastAPI, Google Gemini LLM, and a custom MCP server, enabling intelligent, context-aware conversations for simulated banking operations.
• •Engineered a full-stack architecture with REST and WebSocket APIs, JWT-based authentication, multilingual and accessibility support, and real-time conversational UI built with HTML/CSS/JS.
• •Implemented multi-turn dialogue management with slot filling, context retention, and intent routing, enabling seamless handling of complex banking flows like loan applications, card blocking, and account queries.
• •Integrated LLM-driven NLU (Gemini 1.5 Flash) for intent detection and dynamic response generation, achieving robust conversational understanding in a hackathon-ready, extensible banking assistant system.

Flutter, Node.js, Express.js, Firebase, Machine Learning

• •Fabricated an AI-powered software platform for smart learning and career guidance, with support for 16+ Indian languages.
• •Integrated Gemini-based AI content generation, user surveys, and personalized job dashboards.
• •Implemented a Chrome Extension for auto job applications and AI mock interviews, cutting application time by 50% and increasing user interview readiness.
• •Personalized communities, alumni connect, reward coin system and real-time social feed, fostering 40% higher career networking interactions.

React, Node.js, Express.js, C++, Python, Firebase Functions

• •Extreme-scale, ultra-low-latency hybrid search platform achieving billion-scale search with sub-150 ms latency.
• •Combines sparse search (BM25 + Block-Max WAND) with dense ANN search, orchestrated by learned query budgeter for optimal execution paths.
• •Features hot-reload mechanism for real-time data freshness within 1-2 seconds and scalable C++ core with horizontal sharding.
• •Implements sparse-first routing strategy for cost-efficiency and fast sparse queries (≤80 ms) with semantic dense retrieval.

Flutter, Riverpod 2.0, Routemaster, Firebase

• •Developed a mental wellness software application enabling private health data security and user-centric design, supporting personalized care workflows.
• •Engineered automated health reminders and doctor-patient chat modules, improving mental health follow-up adherence by 30%.
• •Introduced community features, daily mental health surveys, and real-time updates backed by natural language sentiment analysis, with a strong focus on security to enhance user experience.

Flutter, Riverpod 2.0, Routemaster, Firebase

• •Created a real-time ambulance tracking software application that reduced emergency response times by 35% using GPS-based live updates and location sharing with hospitals.
• •Handled 50+ simultaneous hospital-ambulance connections with stable performance, ensuring data availability even during peak loads.
• •Used Firestore and efficient data handling to maintain stable performance during peak usage and ensure quick updates across devices.
• •Integrated features like traffic-aware routing, estimated time of arrival (ETA), and nearest hospital detection to assist in faster decision-making.