
I’m an AI-focused Backend Engineer specializing in LLM-powered applications, RAG pipelines, and intelligent systems, with nearly 3+ years of hands-on experience operationalizing AI solutions in real-world, production environments.
I have designed and deployed 15+ AI-powered APIs and intelligent services leveraging Large Language Models, embeddings, semantic search, and Retrieval-Augmented Generation (RAG). My work includes multi-agent workflows, document-aware chatbots, real-time AI inference, and performance-optimized pipelines that improve accuracy, latency, and scalability.
These AI systems are backed by a strong Python & Django-based backend foundation. I’ve engineered 60+ RESTful APIs and microservices using Django, Django REST Framework, FastAPI, and Flask, enabling real-time communication with WebSockets, background processing with Celery, and scalable data workflows using Redis, PostgreSQL, and vector databases such as Pinecone, FAISS, Weaviate, and Chroma.
Currently, I work on cloud-native AI platforms using Docker, CI/CD pipelines, and Google Cloud Platform (GCP), focusing on system reliability, monitoring, and responsible AI practices. My goal is to build AI solutions that are not only intelligent, but also secure, maintainable, and production-ready at scale.
Tech Stack
APIs Delivered
Years Experience
AI Powered Solutions
Here are a few projects I've worked on recently.

The AI-powered Quran learning platform is an intelligent system designed to support learners of all levels by providing real-time feedback and guidance during recitation. Using advanced speech recognition and natural language processing, the platform listens closely to the user’s recitation and immediately detects mistakes such as mispronunciations, missing letters, incorrect elongations, and Tajweed rule violations. Whenever an error occurs, the system highlights the specific word or verse, demonstrates the correct pronunciation and articulation, and encourages the learner to repeat until they achieve accuracy. This interactive correction process helps learners build confidence, fluency, and proper Tajweed application, ensuring that they are not just memorizing the Quran but reciting it with excellence. Beyond error detection, the platform provides progress tracking, personalized learning paths, and pronunciation improvement tips tailored to each learner’s pace, and consistent feedback.

Real-time transcription is the process of converting spoken words into text instantly as someone is speaking, making it especially useful in lectures, conferences, or live events. Recording transcription, on the other hand, is done after a meeting or event has ended, where the audio or video recording is processed to generate accurate text. Meeting transcription specifically refers to capturing discussions during business or team meetings, often including speaker identification and highlighting key points or action items. Live transcription is very similar to real-time transcription but usually emphasizes displaying captions immediately for participants, such as subtitles during a presentation or webinar. Teams meeting transcription, like in Microsoft Teams or Zoom, combines these approaches by providing both real-time captions during the meeting and recorded transcripts afterward, ensuring accessibility, searchable records, and automated note-taking for participants.

By integrating the GPT API with a plant disease detection model, the system goes beyond simple classification to deliver a comprehensive, AI-driven plant health assistant. The detection model analyzes leaf images to identify whether a plant is healthy or affected by a specific disease. Once the disease is detected, the GPT API transforms the technical results into clear, human-friendly explanations. These responses include the name and description of the disease, its possible causes, early warning signs, recommended treatments or cures, preventive measures, and best practices for maintaining plant health. Additionally, the system offers guidance on optimal growth conditions such as temperature, sunlight, soil quality, and watering routines, ensuring that farmers and gardeners receive holistic care recommendations rather than just a diagnosis.
The Hair Product Recommender is an innovative AI-powered system designed to simplify the way people choose hair care products. By combining QR code scanning with manual input, it allows users to instantly verify whether a product suits their unique hair type and concerns. Through the QR code, the system analyzes the product’s ingredients and formulation, while manual input enables users to share details such as hair type, scalp condition, lifestyle, and climate factors. Powered by a machine learning matching engine, it cross-checks this information with dermatology datasets, product databases, and customer reviews to deliver evidence-based recommendations. The system also integrates context-aware intelligence by factoring in real-world conditions like humidity, pollution, and UV exposure, ensuring that suggestions are practical and personalized.

The Face Recognition and Emotion Detection system is an intelligent AI solution that combines computer vision with OpenAI’s natural language capabilities. Using deep learning models, the system detects faces, recognizes individual identities, and classifies emotions such as happiness, sadness, anger, or surprise in real time through live video streams. Once an emotion is detected, OpenAI is integrated to generate meaningful, human-friendly insights, explaining what that emotion may represent in behavioral or situational contexts. This makes the system not only a powerful tool for security and recognition but also a smart assistant for understanding human emotions in applications like healthcare, education, retail, and human-computer interaction.

The Python Django JWT Authentication System is meticulously designed to ensure robust and secure user access. It seamlessly integrates Django's powerful framework with JSON Web Tokens (JWT) for authentication, offering comprehensive features such as user sign-up, sign-in, and account activation. Upon registering, users receive an email containing a secure activation link, enabling them to activate their accounts safely and effectively. This approach not only enhances user experience by streamlining the activation process.

The Health Dental Care leverages advanced technologies like LLM2, ChatGPT-3.5, and Python to provide comprehensive and intelligent dental care solutions. Utilizing generative AI, the system responds to patient queries with accurate and personalized information. Patients can input their dental health questions and receive immediate, expert-level responses, enhancing their understanding and management of dental issues. This innovative approach ensures timely and informed dental care, improving patient satisfaction and health outcomes through the power of AI-driven interactions.

The Travel Website API is a powerful and comprehensive solution designed to facilitate seamless travel planning and booking. It integrates various travel services, providing users with access to flights, hotels, car rentals, and other travel-related resources. The API is designed to be highly scalable and secure, ensuring a smooth experience for users and developers alike. Key features include token-based authentication, OTP verification, and email confirmation for enhanced security.

Paint Job Website is a robust software solution tailored for painting businesses, enabling seamless management of clients, technicians, properties, jobs, services, and payments. It features a user-friendly interface to add and track clients, schedule and monitor jobs, manage technician assignments, and generate daily job summaries. Utilizing a Redis server for efficient data storage and real-time updates, the software streamlines operations and enhances business efficiency, making it an essential tool for painting companies aiming to optimize their workflow and improve client satisfaction.

All Noor Hotel Skardu is a premier accommodation option nestled in the breathtaking landscapes of Skardu, offering stunning views of the majestic Karakoram Range. Known for its blend of modern amenities and traditional hospitality, the hotel ensures a luxurious and comfortable stay for its guests. Each room is elegantly furnished, providing a cozy retreat after a day of exploring the natural beauty of the region. The hotel’s online booking system allows guests to check room availability and make reservations effortlessly from anywhere in the world, ensuring a seamless planning experience. Whether you're planning a vacation or a business trip, All Noor Hotel Skardu caters to all your needs with top-notch service and facilities. Enjoy hassle-free check-in and check-out processes, designed to make your stay as convenient as possible. Experience the perfect blend of comfort, convenience, and scenic beauty at All Noor Hotel Skardu.

The OCC Manufacturing Backend is a platform designed to manufacturing process with enhanced security measures and user-friendly features. Central to its functionality is token-based authentication. The system allows users to easily upload STP (Standard for the Exchange of Product model data) files and efficiently handles these files by extracting essential dimensions such as length, width, height, area, and density. Once the STP files are uploaded, the system automatically extracts key dimensions, providing users with the flexibility to adjust these sizes to meet specific manufacturing requirements.The design product can order from diffent countries.

The OCC Manufacturing Backend is a platform designed to manufacturing process with enhanced security measures and user-friendly features. Central to its functionality is token-based authentication. The system allows users to easily upload STP (Standard for the Exchange of Product model data) files and efficiently handles these files by extracting essential dimensions such as length, width, height, area, and density. Once the STP files are uploaded, the system automatically extracts key dimensions, providing users with the flexibility to adjust these sizes to meet specific manufacturing requirements.The design product can order from diffent countries.

The Python Django JWT Authentication and Stripe Payments System is a comprehensive solution for secure user authentication and seamless payment processing within web applications. Leveraging Django's robust framework and integrating with Stripe's payment platform, this project offers a seamless user experience while prioritizing data security and privacy. Upon successful transactions, users receive email confirmations, providing them with receipts. Additionally, users can email receipts to themselves or others for record-keeping purposes.