AqsA Baig

AQSA ZAM ZAM MIRZA JOHAR BAIG — Portfolio & Projects

As an AI/ML engineer and full-stack developer studying at VIIT Pune and IIT Madras, I build production-grade applications. Below are my featured projects — each deployed, functional, and impacting real users.

Mahalaxmi Tailors

Production-ready e-commerce platform for a tailoring business. Features secure RBAC, JWT auth, Razorpay integration, and automated AWS deployment using CloudFormation.

MERNAWSRazorpayJWTDevOps

FalcoVita

Scalable healthcare platform with asynchronous task pipelines using Celery/Redis. Includes 20+ interactive data visualizations and multi-layer cryptographic security.

Vue.jsFlaskRedisCeleryOpenAI

IPO-Success-Predictor

Machine learning application achieving 80% prediction accuracy using Ensemble Learning. Deployed on Hugging Face with an interactive assessment interface.

PythonEnsemble LearningHugging FacePandas

My Development Philosophy

Every project in this portfolio started with a real problem. AQSA ZAM ZAM MIRZA JOHAR BAIG believes that the best software is not the most technically complex — it is the software that solves a problem clearly, performs reliably in production, and can be understood and extended by future developers. This philosophy drives every architectural decision, from choosing the right database schema to structuring API endpoints.

For Mahalaxmi Tailors, the goal was building a complete e-commerce system that a non-technical business owner could rely on daily. That meant prioritizing uptime with AWS, keeping payment flows trustworthy with Razorpay webhook verification, and making sure every order notification reached the right person instantly. The system went live within three days of conception and has handled real customer transactions.

FalcoVita tackled the healthcare domain — where data accuracy and security are non-negotiable. The asynchronous architecture using Celery and Redis ensured that heavy ML inference tasks never blocked the user interface. Cryptographic data protection was implemented at the field level, not just at the transport layer.

The IPO Success Predictor demonstrates a complete machine learning workflow: feature engineering from financial datasets, training an ensemble model (80% accuracy), exporting it, and deploying an interactive interface on Hugging Face — making ML accessible to non-developers.

Tech Stack Expertise

Frontend

Next.js 16 (App Router, Server Components, Server Actions), React 19, Vue.js 3 (Composition API), Tailwind CSS, Framer Motion, Chart.js for data visualization.

Backend

Node.js with Express.js for REST APIs, Flask (Python) for ML-integrated backends, Celery + Redis for async task queues, JWT authentication with RBAC, Razorpay and Cloudinary integrations.

Cloud & DevOps

AWS EC2, S3, RDS, Lambda, CloudFront, CloudFormation (Infrastructure as Code), IAM roles and policies, GitHub Actions CI/CD pipelines for automated deployments.

AI & Data

Python (Pandas, NumPy, Scikit-learn), Ensemble Learning (Random Forest, Gradient Boosting), OpenAI API integration, Hugging Face Spaces deployment, data preprocessing and feature engineering.

Read the Full Build Stories

Deep-dive technical articles on how each project was designed, built, and deployed by AQSA ZAM ZAM MIRZA JOHAR BAIG.