Index

KPRY is an AI-powered aggregator built for Pakistan's local streetwear and fashion ecosystem. Instead of jumping between different websites and DMing Instagram pages for prices, KPRY lets you search the entire local market in one place. real-time, one screen.

Type what you want. It finds it across every local brand instantly.

Built to solve a real gap: Pakistan's streetwear scene is growing fast, but the online shopping experience hasn't kept up. KPRY bridges that, covering everything from mainstream labels to underground Instagram-only brands, giving both shoppers and brands a better way to connect.

Architected and deployed a containerized, full-stack data aggregator on a Linux VM using Docker and NGINX, showcasing advanced data pipelines, edge security, and resilient DevOps workflows.

• Engineered resilient data ingestion pipelines using Python, Playwright, and Telethon to systematically bypass strict WAFs and extract dynamic content from tier-1 news syndicates.
• Containerized the application stack (Django, PostgreSQL, scraper loop) with Docker Compose, enabling isolated, reproducible, and portable multi-service deployments.
• Enforced secure edge routing by configuring an NGINX reverse proxy integrated with Cloudflare DDoS mitigation and Certbot for "Full (Strict)" end-to-end SSL encryption.
• Parameterized environment configurations via injected .env variables, isolating PostgreSQL traffic entirely to internal container networks for a zero external attack surface.
• Administered host-level networking and resolved critical Docker iptables state corruption.
• Optimized Layer 7 traffic routing by configuring Django proxy headers and serving static assets blazingly fast via WhiteNoise.

Designed, containerized, and deployed a full‑stack movie recommendation system emphasizing automated delivery pipelines, scalability, and production‑grade DevOps workflows.

• Built the stack with FastAPI (backend) and Streamlit (frontend), exposing a clean REST API for interactive recommendations.
• Containerized services with Docker and orchestrated them using Docker Compose for multi‑service deployments on Azure Linux VMs.
• Implemented a CI/CD pipeline in GitHub Actions that automatically triggers on pushes, uses secure SSH key injection, and executes zero‑downtime redeployments.
• Achieved consistent builds, reproducible deployments, and fully automated delivery from commit to running containers.

Engineered and automated cloud-native infrastructure on AWS using Terraform, showcasing production-grade IaC, immutable provisioning, and secure DevOps workflows.

• Authored declarative Terraform modules to provision Ubuntu-based EC2 instances dynamically sourcing Canonical AMIs via AWS data lookups.
• Parameterized deployments with variables for reusable, multi-environment pipelines.
• Enforced secure connectivity by codifying VPC security groups with principle of least privilege ingress/egress policies.
• Practiced advanced IaC lifecycles: drift detection, idempotent rollout, and full teardown.
• Embedded DevSecOps rigor by codifying infra immutably, reducing manual intervention, and ensuring reproducible builds.

Engineered and deployed a production-grade cloud application for secure file uploads to Amazon S3, leveraging modern DevOps workflows, scalable architecture, and enterprise-grade deployment practices.

• Developed backend with Flask and Boto3, implementing secure and efficient file uploads with proper validation and object management.
• Configured Gunicorn as a multi-worker WSGI server and Nginx as a reverse proxy to ensure high concurrency, low latency, and fault tolerance.
• Integrated AWS best practices: secure S3 access, IAM role usage, public/private bucket policies, and presigned URLs for temporary object access.
• Practiced end-to-end deployment automation and operational monitoring: logs via systemd, Nginx optimization, process management, and reproducible builds.

Engineered a serverless, globally distributed portfolio using AWS S3 and CloudFront, showcasing immutable infrastructure, CDN-first delivery, and CI/CD automation with GitHub Actions.

• Codified artifact publishing pipelines in GitHub Actions for automated build, asset versioning, and idempotent S3 uploads.
• Enforced CloudFront edge invalidation policies to guarantee deterministic propagation of static assets across 200+ PoPs.
• Implemented principle of least privilege IAM roles to secure pipeline tokens.
• Delivered a zero-ops, CDN-accelerated, globally accessible portfolio, reducing latency by >50% compared to regional hosting.

Developed and containerized a full-stack notes management system as part of my internship with Systems Limited, focusing on scalability, security, and DevOps readiness.

• Built using the MERN stack with a modular architecture and clean REST API design.
• Fully dockerized frontend, backend, and database using Docker and Docker Compose for cross-platform deployment.
• Configured NGINX as a reverse proxy to route traffic securely and efficiently between frontend and backend containers.
• Achieved <1s container startup time in dev mode and consistent build behavior across multiple test systems.

Led the development of an advanced RAG-based intelligent assistant designed to facilitate natural language querying over complex university admission documents.

• Leveraged Groke for efficient extraction and structuring of domain-specific knowledge from unstructured data sources.
• Utilized Quadrant to generate high-quality vector embeddings and build a scalable, searchable vector database for rapid retrieval.
• Architected a robust RAG pipeline integrating LangChain, FAISS, and OpenAI GPT APIs to deliver precise, context-aware responses.
• Achieved 90%+ response relevance in comprehensive evaluation across 50+ diverse user queries.

Developed a full-stack cross-platform charity donation app to connect users with verified programs and streamline the donation process.

• Built with React Native and Firebase, achieving 100% real-time sync across all screens and <800ms average screen load time.
• Integrated Firebase Auth with secure login/registration, supporting 3+ sign-in methods and session management.
• Designed and implemented a donation flow that led to 90% task completion in user testing.
• Achieved 0 crash reports across multiple devices (Android/iOS) in end-to-end testing.

As part of a Digital Logic and Design course project, we successfully designed and implemented a 4-bit Arithmetic Logic Unit (ALU) capable of performing both basic and complex operations. The ALU supports functionalities such as addition, subtraction, equality and inequality checking, bitwise AND, OR, NOT, XOR, multiplication, division, and other essential logical operations.

We utilized the Arduino Uno microcontroller as the core processing unit to manage inputs and control the operations. We built the logic using discrete ICs including 7408 (AND), 7432 (OR), 7404 (NOT), 7486 (XOR), 7483 (4-Bit Adder), and 74153 (Multiplexer).

The project was built using a combination of breadboards and jumper wires for prototyping, alongside resistor arrays and LEDs for input/output visualization.