Yogesht.info
Portfolio
Hands-on projects built to demonstrate real-world architecture patterns. Each project is fully working code โ€” not mockups โ€” designed to showcase production-grade thinking across distributed systems, event-driven design, and AWS.
Projects
โšก
ShopFlow
Event-Driven Microservices
A fully working order processing system built with Node.js and AWS (SNS, SQS, DynamoDB). Demonstrates pub/sub fan-out, dead letter queues, atomic stock reservation, event sourcing, and distributed tracing โ€” all patterns used at scale by companies like Uber and Amazon.
Node.js AWS SNS AWS SQS DynamoDB Express EDA Microservices
View Full Details โ†’
๐Ÿค–
RAG Knowledge Assistant
AI / LLM Engineering
A Retrieval-Augmented Generation (RAG) system that lets you chat with your own documents. Upload PDFs or knowledge-base articles, ask questions in plain English, and get accurate answers grounded in your data โ€” not hallucinations. Built with LangChain, ChromaDB, and a local LLM.
LangChain ChromaDB RAG Node.js Llama Vector DB
๐Ÿšง  Coming Soon
๐Ÿ“Š
Real-Time Analytics Dashboard
WebSocket ยท React ยท Node.js
A live metrics dashboard that streams real-time data from a Node.js backend to a React frontend using WebSockets. Demonstrates high-frequency data ingestion, time-series aggregation, and efficient UI rendering patterns โ€” commonly asked in senior frontend and full-stack interviews.
React Node.js WebSocket Chart.js Redis Real-Time
๐Ÿšง  Coming Soon
EDA Patterns Demonstrated
๐Ÿ“ก
Pub / Sub
SNS fan-out delivers one event to all subscribers simultaneously
๐Ÿ“ฌ
Message Queue
SQS decouples producers from consumers โ€” services run independently
๐Ÿ”
Dead Letter Queue
Failed messages are captured after 3 retries โ€” nothing is silently lost
๐Ÿ“–
Event Sourcing
Immutable audit log records every event โ€” full order history preserved
โš›๏ธ
Atomic Operations
DynamoDB conditional writes prevent overselling under concurrent load
๐Ÿ”—
Correlation IDs
orderId threads through every event enabling end-to-end tracing