Mati Ur Rehman

PROFESSIONAL PROFILE Software Engineering graduate with a strong competitive-programming record. Multiple first-place finishes, ICPC regional gold (3rd in Pakistan), and top finishes in national speed contests. Built projects include a C++ search engine, Dockerized storage/streaming services on GCP, and an ESP32-based irrigation controller. Open-source contributor to DORA-RS, implemented an Arrow::FFI bridge between C++ and Rust. Experienced with Git and cloud-native microservices. Machine-learning and Deep learning Practitioner. Led a team to 1st in the UNOOSA region of the Kibo Robot Programming Challenge and built a YOLOv5-nano detection pipeline plus an Android controller being prepared for the ISS final. EDUCATION Software Engineering SEECS , Islamabad , 3.47 (2026) INTERNSHIP EXPERIENCE TUKL Research and Development Lab 02-Jun-2024 - 08-Sep-2024 Understanding basics of DL using coursework. Identifying research topic related to neural architecture search. Performing literature review. Implementing the proposed architectures and evaluting them. FINAL YEAR PROJECT Disrupting Deepfakes through Adversarial Peturbations Objective: Developed a defensive framework to protect digital identity by proactively "poisoning" images against malicious Deepfake manipulation. Technical Stack: Engineered the entire pipeline using PyTorch, focusing on the generation of adversarial perturbations that are imperceptible to humans but catastrophic for generative AI models. Attack Methodologies: Implemented and optimized multiple adversarial attack vectors, including Projected Gradient Descent (PGD) and Carlini & Wagner (C&W), specifically adapted for the constraints of proactive image defense. Target Architectures: Successfully demonstrated the disruption of state-of-the-art diffusion and image-to-image translation models, including FLUX, Kontext Dev, and InstructPix2Pix. TECHNICAL EXPERTISE OpenCV Made a project which was focused on Object dectection. Used yolov5 nano on edge device for detection. C++ Did a project for implementing the relevant article search engine using c++. Also lot of competitive programming experience in cpp. Python Used python for ML/DL related research projects. Databases Used both sql and nosql databases in projects.

Software Engineering graduate with a strong competitive-programming record. Multiple first-place finishes, ICPC regional gold (3rd in Pakistan), and top finishes in national speed contests. Built projects include a C++ search engine, Dockerized storage/streaming services on GCP, and an ESP32-based irrigation controller. Open-source contributor to DORA-RS, implemented an Arrow::FFI bridge between C++ and Rust. Experienced with Git and cloud-native microservices. Machine-learning and Deep learning Practitioner. Led a team to 1st in the UNOOSA region of the Kibo Robot Programming Challenge and built a YOLOv5-nano detection pipeline plus an Android controller being prepared for the ISS final.