a
ayesha_shahid17

AYESHA SHAHID

@ayesha_shahid17

I build RAG apps ML APIs you can test before you buy live demos included

Pakistan
English, Urdu
About me
Got data, documents, or a model idea? I built the working system, not just a prototype. I specialize in RAG applications (LangChain, FAISS, LLMs), ML prediction APIs (FastAPI, Docker, XGBoost, BiLSTM), and clinical AI systems. Everything I deliver is clean, documented, and production-ready with a live HuggingFace demo you can click before hiring me. Published ML researcher, Springer Nature (2026), first-author clinical prediction system with 96% accuracy. That same rigor goes into every client project. I'm the right fit if you want an ML system that actually works in the real world.... Read more

Skills

a
ayesha_shahid17
AYESHA SHAHID
Offline • 

See my services

AI Integrations
I will build a custom ai chatbot to chat with your pdfs, docs, or knowledge base
From$90 / projectMore details
Review & Rewrite
I will edit and strengthen your ml or ai research paper or lit review
From$20 / projectMore details

Work experience

Self_Employed

Machine Learning Researcher

Self Employed • Self-employed

Dec 2023 - Present • 2 yrs 8 mos

Conducted supervised machine learning and deep learning research for clinical prediction and decision support systems across multiple medical domains, including diabetes, brain tumor detection, ASD classification, and cardiovascular disease risk. Published first-author research in the International Journal of Diabetes in Developing Countries (Springer Nature, 2026). Developed a stacked BiLSTM Decision Support System achieving 96% accuracy and 100% sensitivity on an independent clinical test set, outperforming all traditional ML and deep learning baselines. Built and deployed production ML systems, including a Diabetes Risk Prediction API (XGBoost + Random Forest ensemble, FastAPI, Docker) and a Medical RAG Assistant (LangChain, FAISS, Llama-3.3-70B, Streamlit), both live on HuggingFace Spaces. Trained and evaluated 15+ model architectures, including BiLSTM, LSTM, CNN-LSTM Hybrid, XGBoost, Random Forest, SVM, KNN, VGG16, ResNet50, and Explainable Boosting Machines across medical imaging, NLP, and tabular domains.