I'm Hadi, a Computer Science student in Islamabad specializing in machine learning, deep learning, and applied AI engineering from multi-agent pipelines to production-ready interfaces.
I'm a BS Computer Science student specializing in machine learning, deep learning, and Python-based AI application development. I build GenAI-powered assistants, image classification models, and multi-agent data analysis systems with a strong foundation in data manipulation, statistics, and practical software engineering workflows.
Alongside coursework, I've completed a full-stack web development internship and a series of self-directed AI/ML projects, giving me experience across the full lifecycle of a project from data cleaning and model training to shipping a usable interface.
Certifications
Exploratory data analysis is repetitive and time-consuming when done manually for every new dataset.
Designed a 4-agent pipeline — cleaning, analysis, visualization, and reporting orchestrated with LangGraph, using Llama-3 via the Groq API to generate natural-language insights on top of Pandas/Matplotlib visualizations.
A Streamlit tool that automates the full EDA workflow end to end, from raw data to a written report.
Binary image classification needed a lightweight model that still performs well on limited training data.
Trained a MobileNetV2 transfer-learning model on the TensorFlow cats_vs_dogs dataset at 224×224 resolution, then built an image upload and preprocessing pipeline for real-time inference.
A model that predicts cat or dog in real time from any user-submitted photo.
Wanted a hands-free way to interact with the web, media, and news without touching a keyboard.
Built a Python voice assistant using speech recognition and the OpenAI API to open websites and answer spoken questions, plus music playback from a custom library and NewsAPI headline retrieval.
A fully voice-driven assistant with text-to-speech responses for natural, hands-free interaction.
Needed a lightweight, privacy-respecting chatbot without the overhead of persistent storage.
Built a Streamlit-based chatbot powered by the Google Gemini 1.5 Flash API, with session-based memory only.
A fast, real-time conversational chatbot that keeps no persistent record of user data.
Wanted hands-on practice replicating a real, complex streaming-service interface using only core front-end fundamentals.
Rebuilt the Netflix landing and browse experience with plain HTML, CSS, and JavaScript including a responsive layout, hero banner, and horizontally scrollable content rows.
A pixel-conscious static clone that demonstrates layout, responsiveness, and DOM/JS fundamentals without any framework.
Open to AI Engineer and Data Scientist roles. Reach out directly or connect through any of the channels listed here.