Hello, I'm Harsh Vardhan Singh

Project Overview

A scalable Video Transcoding Service built with Go that leverages AWS services for distributed video processing. The system accepts video uploads, processes them into multiple formats, and makes them available for streaming.

Key Features

• Architected a scalable Video Transcoding Service in Go, utilizing AWS Lambdas, EC2, ECS and S3
• Implemented routing, middleware, and HTTP handling using Gin web framework
• Designed Lambda functions to trigger events upon S3 uploads
• Utilized EC2 instances as workers to continuously monitor the Redis queue for new metadata
• Deployed ECS to spin up containers concurrently, transcode videos using FFmpeg, and upload results to S3

Technical Challenges

The main challenges included managing distributed video processing, ensuring fault tolerance, and optimizing the transcoding pipeline for speed and quality. Used Redis for queueing jobs and tracking progress, while leveraging AWS services for scalability.

Project Overview

A real-time chat application backend built with Go that enables secure communication between users. The system includes authentication, real-time messaging, and persistent storage for message history.

Key Features

• Architected and developed the complete backend infrastructure for a real-time Chat Application
• Leveraged Gin, a high-performance HTTP web framework, for handling routing and middleware
• Implemented JWT (JSON Web Tokens) for secure user authentication and authorization
• Integrated bcrypt for robust password hashing
• Utilized WebSockets to facilitate real-time communication features
• Designed and integrated MongoDB to efficiently manage and store user profiles and chat data

Technical Challenges

Managing concurrent WebSocket connections at scale was the primary challenge. Implemented a connection pool and heartbeat system to maintain connection status. Used MongoDB for storing message history and user profiles, with efficient querying patterns.

Project Overview

A URL shortening service developed in Go that creates customizable short links for long URLs. The system includes Redis caching for improved performance and MongoDB for persistent storage of URL mappings.

Key Features

• Developed a URL shortener service in Go, featuring customizable short IDs
• Implemented routing, middleware, and HTTP handling using Gin
• Implemented Redis for caching recent requests, achieving a 40% reduction in server response time
• Enhanced user experience with faster redirections
• Implemented session-based authentication with Redis
• Employed MongoDB to store user data, including short IDs

Technical Challenges

The main challenge was optimizing for high-throughput redirection with minimal latency. Implemented a two-tier caching system with Redis as the primary cache and an in-memory cache for the most frequently accessed URLs, achieving significant performance improvements.