I am a senior research scientist with over 15 years of experience in different areas of computer vision. I am a self-directed problem solver with the ability to research, develop and implement prototypes. Starting my career in 2006, I have worked in satellite image processing, optical target detection and tracking, computational models of vision, augmented reality, robotic manipulation and language-vision models. My analytical skills are applicable in many industries.One of the biggest challenges in my life was moving with my family to Canada. There were many struggles - being a new father, cold harsh winters, high cost of living, being alone in a new city, and the uncertainty of the future. I was still working on my PhD when we arrived for my contract position at York University. The pressure was immense but I worked daily on writing my thesis (and changing diapers) so I could graduate, find a new job and buy a new home. I am always grateful to Prof John Tsotsos and my York University colleagues for their kindness and support.My work experience in the last 3 years has primarily been in computer vision, robotics and machine learning. 1. Language Vision Models: I have an understanding of some of the state-of-the-art models. I conceptualised a system for Interactive Machine Learning (IML) that achieves user intent in a vision task through two-way interaction. A small part of this IML concept was demonstrated using gradio and HuggingFace libraries (Python). 2. Robotic Manipulation with Reinforcement Learning (RL):I coordinated a small team of researchers in a project for robotic insertion with a real Franka Emika Panda. We implemented, evaluated and improved an insertion agent using state-of-the-art techniques.3. Robotic Manipulation:I had a strong involvement in a 3-year robotics research collaboration with a leading US university. I ensured that requirements, progress and risk were communicated to management. Furthermore, I analysed and evaluated the delivered prototype (C++/Python). This included configuring, installing, debugging, evaluating and analysing the full systems in both simulation (VREP) and real (Epson C4L robot). The risks, limitations and issues related to the technologies were communicated to senior management for future planning. 4. Grasp Pose EstimationI implemented a C++ vision module for grasp pose estimation in point cloud. The application was bin picking in industry.
Listed skills include Image Processing, Computer Vision, Pattern Recognition, Digital Image Processing, and 16 others.
