Muhammad Ghous

AI Engineer with hands-on experience building and deploying end-to-end Machine Learning, NLP, and Generative AI systems across internships and independent projects. Proficient in Python, PyTorch, Hugging Face Transformers, LangChain, LangGraph, FastAPI, and Streamlit. Experienced in building Large Language Model (LLM) applications, agentic AI workflows, Generative AI pipelines, and RAG-based systems for production use.

Skilled in the full ML lifecycle — from data preprocessing and model training to API development, model deployment, and monitoring. Seeking an AI Engineer role (onsite Karachi or remote) to contribute to production AI systems.

Trainee AI Engineer at Wenawa (2025-10 – 2025-12)

• Built an advanced OCR pipeline using PaddleOCR and deep learning models for high-accuracy text extraction from scanned documents and medical reports, achieving multi-model comparison with confidence scoring.
• Integrated BERT-based NLP pipelines for medical entity recognition and semantic search, and exposed functionality via FastAPI with endpoints for single/batch OCR processing and analytics.
• Developed an AI-powered WhatsApp Food Ordering Chatbot using FastAPI, integrating Hugging Face Wav2Vec2 speech-to-text for Urdu/English order processing and OpenStreetMap-based branch routing.
• Built a Streamlit admin dashboard for real-time order management and branch monitoring, with complete WhatsApp Business integration via GREEN-API and webhook handling.

Data Science Intern at AA Solutions (2025-08 – 2025-08)

• Designed and deployed an end-to-end Karachi Air Quality & Health Prediction System integrating Flask RESTful APIs and XGBoost (Bayesian optimization) with real-time data pipelines.
• Built an interactive Streamlit dashboard with predictive analytics, health risk stratification, and environmental trend visualization; implemented microservices architecture and Docker deployment.

Data Science Intern at Digital Empowerment Network (2025-07 – 2025-09)

• Built and deployed an Email Spam Classifier using TF-IDF vectorization and Multinomial Naive Bayes, evaluated with accuracy, precision, and recall metrics; deployed via Streamlit.
• Developed a SpaceX First Stage Landing Prediction system using Random Forest with GridSearchCV, integrating live SpaceX API data with an interactive Plotly & Folium dashboard.
• Built a House Price Predictor using multiple regression models (Linear Regression, Random Forest, Gradient Boosting); selected best model by R² score and MSE after feature engineering and EDA.

Bachelor of Science in Artificial Intelligence – Dawood University of Engineering & Technology (2021-11 – 2025-12)