- 🎓 Incoming MS AI & Data Science Student at GUTECH (Fall 2026) | BS CS from FAST NUCES Karachi
- 💼 AI Research Assistant at GUTECH specializing in Medical AI, Radiomics, and Computer Vision
- 🧠 Passionate about bridging Deep Learning, Medical Imaging, and Computational Neuroscience
- 🎯 Currently exploring the internal geometry of vision models and interpretable AI for high-stakes healthcare.
- 🥉 3rd place in Procom Computer Vision Competition 2025
- 🥉 3rd place in PaysysLab AI Competition 2025
Current
- Radiomics & Survival Prediction: Engineering multi-modal data fusion pipelines for Preoperative Diffuse Glioma MRI (UCSF-PDGM) and local AKU datasets. Implementing advanced feature extraction and statistical modeling in PyTorch to evaluate tumor aggressiveness and predict patient survival (C-index).
- Pediatric Oncology Telemedicine: Designing modular data pipelines using Retrieval-Augmented Generation (RAG) and Milvus to extract grounded clinical insights from unstructured patient reports for AI-driven decision support.
- Fine-Grained Action Recognition: Architecting temporal pipelines combining YOLO, MediaPipe, and Temporal Transformers/LSTMs for highly precise sequence monitoring and anomaly detection (e.g., ATM workflow compliance and industrial assembly).
Project Lead Developed an end-to-end Image-Based Recommender System for Diabetic Foot Ulcer (DFU) treatment.
- Architectures: Achieved state-of-the-art segmentation and classification using a hybrid YOLOv11 + Swin Transformer approach, alongside ConvNeXt and ViT.
- Full-Stack Deployment: Containerized the backend using Docker and FastAPI, creating a robust REST API to generate PDF reports with segmented overlays.
- Clinical Dashboard: Integrated real-time inference into a React Native dashboard to reduce diagnosis turnaround time for clinicians.
- Languages: Python, Node.js, C/C++
- ML & Deep Learning: PyTorch, TensorFlow, Scikit-learn, XGBoost, HuggingFace Transformers
- Computer Vision: Radiomics, YOLOv11, Swin Transformer, ViT, MediaPipe, PatchCore
- GenAI & NLP: RAG Pipelines, Milvus, Vector DBs, Prompt Engineering
- Backend & DevOps: FastAPI (Async), Docker, PostgreSQL, MongoDB, Git
- Advance state-of-the-art research in Medical AI, Neuro-technology, and Human Activity Recognition.
- Probe the representational geometry of vision models to build trustworthy, interpretable AI systems.
- Continue competing and collaborating in national & international AI/Computer Vision competitions.
- LinkedIn: khalid-khurshid-siddiqui-b0b827238
- Email: khalid20031016@gmail.com