Yousaf
Building the future, one microservice at a time.
Overview
Social Links
About
Hey, I'm Muhammad Yousaf — a Full Stack Developer who loves turning complex problems into clean, scalable solutions.
I work primarily with the MERN stack and Next.js, building production-ready applications that balance performance with great user experience. Whether it's architecting REST APIs, optimizing database queries, or translating designs into pixel-perfect interfaces, I enjoy the entire process of bringing ideas to life.
Right now, I'm diving deep into AI development — exploring Agentic AI, LangChain, and LangGraph. I'm fascinated by the challenge of building systems that don't just execute commands but actually think and adapt. Combining my full-stack foundation with AI is where I see the real potential to create something impactful.
Beyond the code, I'm always learning. System design, new frameworks, better patterns — I believe there's always a smarter way to solve a problem, and I'm constantly looking for it.
If you're working on something interesting or just want to talk tech, feel free to reach out. I'm always up for good conversations and new opportunities.
GitHub Contributions
Stack
Experience
Code Expert
A production-level multi-vendor food delivery and gift mart platform with comprehensive role-based access control and module management.
Project Focus:
- Dashboard Architecture: Built two distinct dashboards (Customer, Vendor) with dynamic UI rendering, conditional logic, and real-time data updates
- Authentication & Security: Implemented secure authentication using NextAuth.js with protected routes, session management, and credential handling
- API Development: Designed and integrated RESTful APIs handling orders, food management, and user operations with optimized data flow
- Design Implementation: Translated Figma designs into responsive, pixel-perfect UI components using Tailwind CSS
- Email Integration: Integrated Resend for transactional notifications including order confirmations and status updates
- Database Optimization: Optimized MongoDB queries using Mongoose lean queries and schema design patterns
- Next.js
- Express.js
- Tailwind CSS
- Next Auth
- MongoDB
- ShadCN UI
- Node.js
- Resend
- Figma to Code
- Responsive Design
- API Development
- Teamwork
- Research
- Problem-solving
Projects(4)
An AI-powered sentiment analysis platform that processes YouTube video comments and transcripts to generate actionable insights. Co-developed with 80% backend contribution.
Key Technical Achievements:
- Asynchronous Job Processing: Architected queue-based system using BullMQ and Redis where sentiment analysis requests generate job IDs and are queued for processing by available workers, preventing server overload and ensuring optimal resource utilization
- LangGraph Workflow Engine: Built multi-stage AI workflow with parallel processing pipelines featuring sentiment classification, emotional analysis, content pattern detection, and actionable insights generation across 7 distinct stages
- API Key Rotation & Load Balancing: Implemented intelligent rotation across 8 Google Gemini API keys with automatic load distribution to optimize throughput and prevent rate limit bottlenecks
- Real-time Progress Tracking: Integrated WebSocket communication providing live progress updates through stages: queued → fetching comments → transcript retrieval → classification → parallel analysis → summarization → completion
- Parallel Data Fetching: Optimized data retrieval using Promise.allSettled for concurrent YouTube comments and transcript fetching with graceful fallback handling
- Redis-Based Rate Limiting: Applied feature-specific throttling with express-rate-limit and Redis store to ensure system stability under high concurrent load
- Production Deployment: Configured backend deployment on Render, implemented Resend for transactional email notifications, and optimized Next.js frontend with SEO enhancements on Vercel with custom domain integration
- Worker Concurrency Control: Configured BullMQ workers with concurrency of 3 and rate limiting (2 jobs per day) for optimal performance
- Frontend Contributions: Developed UI enhancements, implemented SEO optimization in Next.js, and improved overall user experience (20% frontend contribution)
Technical Architecture:
- Backend processes comments in batches using structured LLM outputs (Zod schemas)
- Parallel analysis branches for emotions, patterns, loved aspects, improvements, and viewer requests
- Automatic transcript availability detection with non-blocking error handling
- Job progress tracking with percentage-based updates and stage-specific messages
- Node.js
- Express.js
- TypeScript
- LangChain
- LangGraph
- BullMQ
- Redis
- Socket.io
- Next.js
- Resend
- Render
- Vercel