32+
Projects
Available for Hire
Open to work
Biography
I build AI that solves real problems.
I’m Muhammad Farooq Shafi, an AI Engineer & Data Scientist from Pakistan with a passion for turning complex data into intelligent, production-ready systems.
Crafting intelligent systems with Machine Learning, NLP and AI.
My work spans the full AI pipeline — from raw data ingestion to model training, SHAP explainability, FastAPI deployment, Docker containerisation, and GitHub Actions CI/CD. I’ve built over 32 production-grade ML projects, including systems that process 50,000+ daily predictions and models achieving 99.2% AUC-ROC.
Location
Pakistan — Remote Worldwide
Status
Open to Work
Work Type
Full-time / Freelance
Languages
Urdu, English
By the Numbers
32+
Projects
40+
ML Models
50K+
Daily Predictions
500+
Contributions
14+
Certifications
3+
Years Experience
Expertise
Building AI Systems That Matter.
With hands-on experience across financial AI, aviation analytics, NLP and large language models — from raw data to production APIs serving 50K+ daily predictions.
Machine & Deep Learning
End-to-end ML/DL pipelines with PyTorch, TensorFlow, scikit-learn. Fraud detection at 99.2% AUC-ROC, classification to neural networks.
NLP & LLMs
Transformers, BERT, fine-tuning with LoRA/DPO. Qwen2.5 tool-match: 18.7% → 79.8% on 112K examples.
RAG & Vector Search
Hybrid BM25 + FAISS retrieval, cross-encoder reranking, Pinecone/Chroma vector stores. p95 latency 4.93ms.
Computer Vision
Real-time image processing, object detection, multimodal AI with OpenCV and PyTorch.
Data Science & BI
Tableau, Power BI, Pandas, SQL — transforming raw datasets into actionable business intelligence.
MLOps & Production AI
FastAPI, Docker, Kubernetes, GitHub Actions CI/CD, MLflow — scalable AI systems at production grade.