Hamza Shafiq

I am a Senior Research Engineer at IK Lab, specializing in developing and maintaining deep learning algorithms for a wide range of projects. My expertise spans computer vision, deep learning, NLP, and AI-based automation, with extensive work in restoration and enhancement, super-resolution, generative AI,security applications, and AI-generated test cases.

During my master’s, I worked on:

1. Restoration and colorization of old images using adversarial networks.
2. Dental image segmentation for automated diagnostics.

Key Projects at IK Lab

• AI for Security & Surveillance: I developed a deep learning-based Vehicle Accident Investigation Reporting System, integrating Detection, Tracking, Segmentation, Restoration, and Text Recognition for hit-and-run detection, license plate recognition, and impact tracking.
• Turbulence Mitigation for Satellite Detection: I designed algorithms to reduce turbulence effects and enhance satellite detection.
• Edge AI Optimization: I worked on optimizing deep learning algorithms for edge devices using TensorRT and quantization techniques, enabling efficient inference on resource-constrained hardware.
• Damage Detection & Localization: I applied deep learning to analyze aerial images, identifying the extent and location of damage for various applications.
• AI-Based Test Generation: I leveraged Codex, CodeBERT, GPT, and LLaMA to generate test cases and compared them with traditional methods using BLEU/ROUGE scores, code coverage, and bug detection.

Technical Skills & Interests

• Deep Learning & Computer Vision: Stable Diffusion, Transformers, GAN, Mamba
• NLP & Test Generation: GPT, LLaMA, Codex, CodeBERT, BLEU/ROUGE evaluation
• AI for Security & Surveillance: Hit-and-run detection, satellite imagery analysis
• Software & System Development: FastAPI, Gradio, OpenCV, PyTorch, TensorFlow, TensorRT, JAX, Docker