Skip to content
View AyaanHussain1's full-sized avatar
  • Student / Aspiring AI Engineer
  • Karachi, Pakistan

Block or report AyaanHussain1

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AyaanHussain1/README.md 

+ Building RAG pipelines that actually cite their sources
+ Shipping ML models past a single training script
+ Turning resumes, transcripts, and raw code into structured data
+ Open to AI/ML Engineer roles

Shipped Models Accuracy

🧭 Start here

A concise overview of the work that best represents my engineering approach:

  1. CodeSage — my flagship build. A RAG engine shipped across three real client surfaces, not a single-file demo. Try the live demo →
  2. Ezitech AI Talent Platform — production engineering work: schema-validated LLM output, not just a prompt wrapper.
  3. ML Model Suite — 15+ models, every one with a reported metric and a runnable .py file.

🧩 Featured Work

CodeSage

A RAG engine, not a wrapper. Query YouTube transcripts and local codebases from a Chrome extension, an Electron desktop app, or the web — all hitting the same retrieval backend.

Deployed: Railway (backend) + Vercel (frontend)

Live demo: rag-agent-frontend-three.vercel.app

AI Talent Discovery Platform

Engineering work @ Ezitech. Parses resumes with pdfplumber + GPT-4o-mini, validates every LLM output against a Pydantic schema, scores portfolios, and matches candidates to jobs on skill overlap.

Status: Built and pushed to GitHub

📦 Expand: Full ML Model Suite (15+ projects, all metric-backed)
Project Approach Result
ASD Screening Classifier XGBoost 93% accuracy
Credit Card Fraud Detection Scikit-learn, imbalanced classification 93% accuracy
SMS Spam Classifier Naive Bayes + NLTK 96.8% accuracy
Car Price Predictor Scikit-learn regression + Streamlit UI R² ≈ 0.74

+15 more model repos on my profile — every one has a runnable .py entry point and a real README, no notebook-only projects.

🛠️ Stack

Python TensorFlow Scikit-learn LangChain FastAPI MySQL Docker Git

📈 Activity

💬 Let's build something

Open to AI/ML Engineer, RAG/LLM Systems, and Backend for AI Products roles.

Popular repositories Loading

  1. CodeSage_Agent CodeSage_Agent Public

    AI agent that answers questions about YouTube tutorials and your own codebase — available as a Chrome extension, desktop app, and web app.

    JavaScript 1

  2. Rag-Agent Rag-Agent Public

    LearnShift AI turns static Java OOP content into a personalized AI tutoring experience that supports both learners and educators.

    Python

  3. Autism-Prediction Autism-Prediction Public

    ML web app predicting Autism Spectrum Disorder traits from AQ-10 screening responses. Built with XGBoost (93% accuracy), Streamlit, and scikit-learn. Supports single and batch predictions.

    Jupyter Notebook

  4. Credit-Card-Fraud-System Credit-Card-Fraud-System Public

    ML web app detecting fraudulent credit card transactions using Logistic Regression (93% accuracy) on a highly imbalanced real-world dataset. Built with scikit-learn and Streamlit, using undersampli…

    Python

  5. Book-Recommendation-System Book-Recommendation-System Public

    Book recommendation web app using item-based collaborative filtering (K-Nearest Neighbors, cosine similarity) on 1.1M+ user ratings. Built with scikit-learn and Streamlit.

    Python

  6. Car-Price-Predictor Car-Price-Predictor Public

    Car price prediction web app using Linear Regression (0.74 CV R²), selected via fair cross-validation over Random Forest and Gradient Boosting. Built with scikit-learn and Flask, with dynamic compa…

    Jupyter Notebook