SaiSunnyT.

- Developed a fully distributed, decentralized content-addressable file storage system, capable of handling and streaming very large files for high availability and data durability. - Built a TCP-based Peer-to-Peer (P2P) networking system with a custom transport layer, enabling decentralized, secure communication between nodes via bi-directional messaging and streaming. - Designed and implemented POST and GET APIs for efficient file storage and retrieval. - Developed a file replication mechanism, distributing files across peer nodes to enhance availability and fault tolerance. - Implemented encryption and decryption to secure data during transmission and encoding/decoding(gob) for optimized communication. - Enabled multi-writer support, allowing multiple nodes to write data concurrently while maintaining consistency. - Introduced a control frame using bit messages to distinguish between messaging and streaming operations. - Utilized waiting groups to efficiently synchronize and manage parallel operations. - Future enhancements will include adding update and delete operations.

Voice AI Agent: Configures a Twilio phone number to forward calls, encoding audio as base64 and sending it to a server via ngrok. ngrok forwards audio to a local Flask application, which greets the caller and handles the call flow using Twiml. Streams audio to a WebSocket route, decodes base64, and sends it to AssemblyAI for transcription using the Mulawa algorithm. AssemblyAI processes audio in real-time, printing partial and final transcripts on the terminal. --- Conversational AI Agent: Eleven Labs generates an audio greeting streamed to the customer. Customer inputs are captured via microphone, transcribed by AssemblyAI. Transcribed text is passed to the Gemini pro model, which generates a response. The response is converted to speech by Eleven Labs and played back to the user. Plans to add features like managing database queries and booking appointments. Future Enhacements: These projects can be extended to improve various functionalities in the future like adding a conversational way, booking, creating events in calendar, and db setup. System Design: https://lucid.app/lucidchart/808b5a91-18ee-4360-a9ba-c90918fffde4/view

Designed and deployed a scalable Three-Tier architecture on AWS EKS, automating infrastructure with Terraform & integrating DevSecOps best practices using OWASP Dependency Check, Trivy, and SonarQube in CI/CD. • Developed Jenkins pipelines for app deployment (Docker, security scans, ECR, ArgoCD) and EKS management (Terraform LifeCycle). Set up Prometheus & Grafana for monitoring and cost optimization using CAST AI. • Established a jump server for secure cluster access and management. Configured Kubernetes networking & security with NGINX Ingress, AWS ALB, IAM, Secrets, and Kyverno. Leveraged ArgoCD for GitOps-driven deployments.

Designed and implemented a real-time ETL data warehouse architecture leveraging Apache Airflow, Kafka, and Apache Pinot to support both batch and streaming data pipelines. Developed custom Apache Airflow hooks to seamlessly integrate Kafka and Apache Pinot, enhancing real-time data ingestion and ETL efficiency. Optimized data visualization and reporting by integrating Apache Pinot with Kafka and building insightful dashboards in Apache Superset, utilizing dimensional modeling for improved data organization and analytics. Deployed project applications, including Apache Airflow, Kafka, Apache Pinot, and Apache Superset, using Docker containers orchestrated via Docker Compose

Developed a full-stack Beat Box Application leveraging Next.js 13.4 and React for a responsive user interface, styled with Tailwind CSS. Utilized Supabase for authentication and PostgreSQL for database management, ensuring secure user registration and login processes. Integrated Stripe to manage premium subscriptions, handling secure transactions and user billing information seamlessly. Implemented advanced audio playback features, allowing users to stream music effortlessly across devices. Designed and built a comprehensive favorites and playlist system to enhance user experience, enabling personalized music collections. Used Supabase storage to manage file and image uploads securely, ensuring smooth and efficient handling of user data. Incorporated react-hook-form for client-side form validation and react-toast for server-side error handling, enhancing application stability and user interaction. Developed route handlers for POST operations, directly accessing the database from server components without the need for API calls. Managed real-time data synchronization between server and client components, ensuring a responsive and up-to-date user experience.

Developed a dynamic Next.js 14 & React application featuring advanced AI integrations and modern styling with Tailwind css. Integrated a chat bot with Vercel AI SDK and enhanced its capabilities through Gemini AI API. Utilized Langchain for various AI functions, including ChatGeminiAI, templates, and sophisticated document processing with DirectoryLoader, TextLoader, and RecursiveCharacterTextSplitter. Implemented vector embeddings and a vector store using AstraDB, and optimized data retrieval with Langchain vector retrievers and retrieval chains. Incorporated chat history management and caching with Upstash Redis for improved performance.

Developed a full-stack web and mobile application using Android Studio with Java for the mobile app and Flask with MySQL for the backend, leveraging RESTful APIs to streamline loan applications, enhance financial management for women beneficiaries, and ensure a seamless user experience across all platforms. Implemented key features such as role-based access control, real-time loan application tracking, and a secure live chat system, improving loan processing efficiency and user engagement.