Muhammad Habeel
AI Engineer · Computer Vision & Multi-Modal Systems
// Professional Summary
Results-driven AI Engineer specialized in Computer Vision, Multi-modal Biometrics, Deep Learning pipelines, and high-performance backend architectures. Experienced in building production-grade verification pipelines combining facial recognition (InsightFace), sub-second vector search (FAISS), and custom anti-spoofing liveness classifiers (EfficientNet-B0) deployed via asynchronous FastAPI services.
// Professional Experience
- Architected and deployed production-ready multi-modal biometric attendance pipelines integrating InsightFace Buffalo_L (512-D embeddings), FAISS vector similarity search, and secondary iris pattern verification.
- Developed and trained a custom EfficientNet-B0 anti-spoofing classifier on a self-collected dataset to detect print attacks, screen replays, and physical masks with minimal latency.
- Engineered scalable, asynchronous FastAPI microservices containerized with Docker, optimized for low-latency edge camera ingestion and GPU compute utilization.
- Integrated multimodal retrieval pipelines with vector databases (FAISS, Chroma) to empower intelligent automation workflows and internal analytical agents.
- Conducted research on volumetric medical MRI scan classification and anomaly segmentation using CNN and Transformer backbones with Grad-CAM visual interpretability heatmaps.
- Designed a conversational NLP triage system for ENT and nasal health guidance leveraging Hugging Face Transformers.
- Trained transfer learning models (MobileNetV2, ResNet50) on 140k+ Kaggle face datasets to classify and detect deepfake synthesis artifacts.
// Selected Engineering Projects
Complete verification system featuring face embeddings search across 10,000+ IDs, iris verification, and real-time anti-spoofing liveness detection.
Full-stack cultural discovery platform with hardened security middleware, rate limiting, and interactive itinerary routing.
Deep learning classification notebook on Kaggle 140k dataset benchmarking residual architectures for synthetic facial boundary artifact detection.
// Technical Competencies
// Education
Specialization in Deep Neural Networks, Computer Vision & Multimodal Systems