Ashad Qureshi@ashad_443

Begin a revolutionary career journey with our AI-driven platform. Our advanced technology extracts data from your resume, offering personalized insights to enhance your professional story. Elevate your readiness for interviews through our simulation feature, which accurately mirrors real-world scenarios. Our platform then utilizes the extracted resume data to conduct and evaluate interviews, ensuring a thorough assessment tailored to your strengths. Discover your untapped potential and master interviews with our all-encompassing approach. Your journey to success commences here, with AI at the heart of career transformation.

Personify AI Chatting Evaluator and Assistant is a sophisticated tool designed to elevate communication on WhatsApp and LinkedIn. It offers real-time message suggestions and engagement scores to help users craft more effective and impactful responses. By integrating seamlessly with both platforms, Personify delivers context-aware replies and insightful feedback tailored to the specific dynamics of each conversation. The tool provides users with the ability to customize message tone and style, ensuring that suggestions align with their personal or professional communication needs. It includes advanced features such as sentiment analysis, which assesses the emotional tone of interactions, and a user feedback system that allows continuous improvement of the AI’s recommendations. By leveraging these capabilities, Personify enhances conversational effectiveness and efficiency, making interactions on WhatsApp and LinkedIn more meaningful and personalized. Whether for casual chats or professional networking, Personify helps users engage more effectively and achieve their communication goals.

Let’s face it, figuring out how to arrange a room is like solving a puzzle—except the pieces are heavy furniture, and you’re not a designer. Whether you’re cramming a couch into a tiny studio or trying to make sense of an awkwardly shaped living room, it’s tough. And unless you’ve got an interior design degree hidden somewhere, most tools out there aren't much help. Enter RoomAligner, your AI-powered, no-stress room layout wizard. We take the heavy lifting off your shoulders (literally, please don’t lift that couch) by combining object detection, natural language processing, and space optimization. In plain English: we’ll spot your furniture in floor plan images, tell you where it is, and give you smart suggestions on how to make everything fit better—like a pro interior designer, but without the bill. The process is a breeze: just upload a room or floor plan image, and boom—our AI swoops in, detects your furniture, breaks down the room’s current setup in plain language, and drops some slick suggestions on how to improve your space. Need to shift that couch or rethink where the TV goes? We’ve got you covered. The magic happens on the backend thanks to FastAPI and Llama models, while the frontend is powered by React Native and Firebase to give you a seamless app experience on any device. Whether you're a design newbie or just trying to avoid another argument about where the coffee table should go, RoomAligner is here to make room optimization fun and easy!

A three-fold architecture to assume diverse roles for LLAMA-3.2B in order to solve the language barrier problem, generate answers in local language (Urdu), and only cater Pakistan's Tax Ordinance. Now we achieved this first by creating a RAG pipeline, where we used mini-LLM to generate embeddings for our documents, we had two docs, 1 referred to 800 pages of laws governing taxation, other referred to different combination of designations requiring different slabs of tax computation. Now we stored chunks + embeddings in a VectorDB and upon query of user, extracted the relevant context through similarity search, after that we used LLAMA-3.2-11B to match the context with the query and create keyword specific tags, that would help streamline the prompt further ahead on basis of user's query and knowledge base. Then we sent those tags + query + context to LLAMA again to generate the relative pointers to each tag in a detailed simplified manner according to the context. Then we sent this generated text to LLAMA acting as a language converter, to prompt for Urdu Answers such that it can be understood by the majority of native population.