Skip to content
View AYESHAASS's full-sized avatar

Block or report AYESHAASS

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
AYESHAASS/README.md

Hi, I'm Ayesha 👋

ML Researcher & AI Engineer | Healthcare AI · RAG · Agentic Systems

I take ML models from research to production: published clinical AI research, 4 live deployed applications, and hands-on experience building RAG and multi-agent systems.

📩 ayesha.psr1234@gmail.com · 🔗 LinkedIn · 📄 Published Paper (Springer Nature)

🔬 Research

First-author publication, Springer Nature, International Journal of Diabetes in Developing Countries (2026) A hybrid BiLSTM clinical decision support system achieving 96% accuracy and 100% sensitivity on independent clinical validation (N=84), outperforming all baseline models (McNemar test, p = 0.0063). 📄 Read the paper

📌 Featured Projects

Multi-Agent Clinical Decision Support System (Google x Kaggle AI Agents Capstone) 4-agent clinical pipeline (Validator, Risk Predictor, Guideline RAG, Clinical Auditor) orchestrated with Google ADK. Wraps an XGBoost ensemble as an MCP Server tool and cross-checks predictions against ADA 2023 / PES 2022 guidelines. 🔗 Live demo

Agentic Clinical Decision Support System (Cotiviti Intern Assessment) 3-layer agentic pipeline: XGBoost + Random Forest ensemble (76%+ accuracy), SHAP explainability, and a Llama-3.3-70B reasoning agent grounded via RAG in ADA 2023 guidelines for auditable, citable outputs.

Medical RAG Assistant Multi-document RAG system with LangChain, FAISS, and Llama-3.3-70B. Recursive semantic chunking, 384-dim HuggingFace embeddings, and an anti-hallucination framework with source citations. 🔗 Live demo

Diabetes Risk Prediction API XGBoost + Random Forest voting ensemble achieving 92% accuracy and 0.88 F1-score, served via FastAPI, containerized with Docker. 🔗 Live demo

🛠️ Tech Stack

AI Specialization: Generative AI, Agentic AI, Multi-Agent Systems, Google ADK, MCP Server, RAG, LLMs, Healthcare AI, Clinical Decision Support Core Models: BiLSTM, XGBoost, Random Forest, Pretrained CNNs (VGG16, ResNet50) LLM & RAG Stack: LangChain, FAISS, Groq API, Llama-3.3-70B, HuggingFace Embeddings, Prompt Engineering Deployment: FastAPI, Docker, Streamlit, HuggingFace Spaces, Uvicorn, Pydantic Languages & Tools: Python, MySQL, Flask, Git, GitHub

🌍 Currently

  • ✍️ Co-authoring a second manuscript, currently in revision
  • 🎓 Research Collaborator at GCWU Sialkot, extending clinical decision-support research
  • 🎓 Completed Google x Kaggle's 5-Day AI Agents Intensive (Agents for Good track)

📫 Reach me: Email · LinkedIn

Popular repositories Loading

  1. diabetes-fastapi-service diabetes-fastapi-service Public

    A production-grade ML Microservice using an Ensemble Model (XGBoost + Random Forest) served via FastAPI and Docker.

    Python

  2. Medical-RAG-Assistant Medical-RAG-Assistant Public

    🩺 A production-grade, hallucination-resistant RAG engine for medical manuscripts. Built with LangChain, Llama-3.3-70B, and FAISS for clinical precision.

    Python

  3. multi-doc-rag multi-doc-rag Public

    Python

  4. Clinical-MultiAgent-CDSS Clinical-MultiAgent-CDSS Public

    A research backed multi agent clinical decision support system for diabetes risk prediction using Google ADK and MCP protocol.

    Python

  5. Agentic-AI-for-Clinical-Decision-Support-System Agentic-AI-for-Clinical-Decision-Support-System Public

    A Clinical Decision Support System (CDSS) for Diabetes Risk using Agentic AI & SHAP Explainability. Developed for Cotiviti Topic 2: Prediction, Pattern Recognition, and Agentic Generative AI for TPO.

    Python

  6. AYESHAASS AYESHAASS Public