Pubudu Sanjeewani

Research Interests

Explainable AIAI for ScienceFoundation ModelsGenerative AINLPComputer VisionHealth & Medical AIMachine Learning

Hello! I'm Pubudu Sanjeewani

Associate Lecturer · RMIT University

I am an Associate Lecturer in Artificial Intelligence at RMIT University, Melbourne. My research interests include deep learning, computer vision, explainable AI, generative AI, machine learning, and artificial intelligence. I hold a PhD in Artificial Intelligence and Image Processing from Central Queensland University, and I have previously worked as a research fellow, machine learning engineer and software engineer.

My current research focuses on the medical and health domain, where I develop methods for analysing multimodal medical data, including images (histopathology, CT, MRI), signals (ECG) and text, with a focus on explainable AI, generative AI, and large language and vision-language models. Earlier in my career, I applied these methods to road safety, detecting safety attributes from roadside video, and to workplace safety.

I am open to supervising PhD and Master's research students.

Education

Ph.D. in Artificial Intelligence and Image Processing

Centre for Intelligent Systems (CIS), Central Queensland University – Brisbane Campus, Australia

Industry partners: Department of Transport and Main Roads (DTMR), Queensland, and the Australian Road Research Board (ARRB).

Thesis: A Deep Learning Technique for the Analysis of Road Safety Attributes.

Supervisors: Prof. Brijesh Verma and Dr Mary Tom.

Bachelor of the Science of Engineering (Hons) in Electrical and Information Engineering

University of Ruhuna, Galle, Sri Lanka (Washington accord accredited)

Employment History

Associate Lecturer

RMIT University – City Campus, Melbourne, Australia

Sessional Academic: Data Science and AI

RMIT University, Melbourne, Australia

Postdoctoral Research Assistant

RMIT University, Melbourne, Australia

Research Fellow (remote)

Griffith University, Brisbane, Australia

Postdoctoral Research Assistant

RMIT University, Melbourne, Australia

Machine Learning Engineer

Smart AI Connect, Melbourne, Australia

Intern

Smart AI Connect, Brisbane, Australia

Consultant

KJR, Melbourne, Australia

Consultant

Datarwe, Brisbane, Australia

PhD Researcher

Department of Transport and Main Roads (DTMR), Queensland, and Australian Road Research Board (ARRB), Australia

Software Engineer

Sierra Construction Limited, Sri Lanka

Software Engineer

Sierra IoT Services (Pvt) Limited, Sri Lanka

Software Engineer

Vogue Tex (Pvt) Ltd, Sri Lanka

Academic Instructor

University of Moratuwa, Sri Lanka

Academic Instructor

University of Ruhuna, Sri Lanka

Trainee Engineer

Ceylon Electricity Board, Sri Lanka

Trainee Engineer

Etisalat Lanka (Private) Limited, Sri Lanka

Publications

2026

  • P. Sanjeewani, B. Verma (2026). A deep learner with small training data for the detection of road safety attributes. Neural Computing and Applications, vol. 38, no. 665. DOI
  • M. A. Rahman, P. Sanjeewani, A. Pham, A. Perera, A. Alavi, U. K. Wiil (2026). Understanding student mental health: An explainable pattern analysis approach. Acta Psychologica. DOI
  • A. Rodrigo, I. Munasinghe, P. Sanjeewani, A. Perera (2026). Benchmarking CNN and transformer-based object detectors for UAV solar panel inspection. 4th IEEE International Conference on Image Processing and Robotics (ICIPRob), Mount Lavinia, Sri Lanka. DOI
  • B. Azam, P. Sanjeewani, B. Verma, A. Rahman, L. Wang (2026). Unified graph-based framework for visual explainability in convolutional neural networks. Information Sciences, vol. 723. DOI

2025

  • H. Nimesha, P. Sanjeewani, A. Perera (2025). Deep ensemble model for fine-grained breast tumor classification across multi-magnification samples. 7th IEEE International Conference on Advancements in Computing (ICAC), Kotte, Sri Lanka. DOI
  • B. Azam, P. Sanjeewani, B. Verma, A. Rahman, L. Wang (2025). A novel graph-based framework for understanding of decision-making process in deep learning models. IEEE Symposium on Trustworthy, Explainable and Responsible Computational Intelligence (CITREx Companion), Trondheim, Norway, pp. 1–5. DOI
  • B. Azam, D. Kuttichira, P. Sanjeewani, B. Verma, A. Rahman, L. Wang (2025). A novel non-iterative training method for CNN classifiers using Gram-Schmidt process. Neural Processing Letters, vol. 57, no. 27. DOI

2024

  • P. Sanjeewani, G. Neuber, J. Fitzgerald, et al. (2024). Realtime personal protective equipment non-compliance recognition on AI edge cameras. Electronics, vol. 13, no. 15, Art. no. 2990 (Special Issue on Pattern Recognition and Image Processing). DOI

2021

  • P. Sanjeewani, B. Verma, J. Affum (2021). Multi-stage deep learning technique for improving traffic signs recognition. 36th IEEE International Conference on Image and Vision Computing New Zealand (IVCNZ), Tauranga, New Zealand, pp. 1–6. DOI
  • P. Sanjeewani, B. Verma, J. Affum (2021). Multi-stage deep learning technique with cascaded classifier for improving turn lanes recognition. IEEE Symposium Series on Computational Intelligence (SSCI), Orlando, Florida, USA, pp. 1–8. DOI
  • P. Sanjeewani, B. Verma (2021). Optimization of fully convolutional network for road safety attribute detection. IEEE Access, vol. 9, pp. 120525–120536. DOI
  • P. Sanjeewani, B. Verma (2021). Single class detection-based deep learning approach for identification of road safety attributes. Neural Computing and Applications, vol. 33, pp. 9691–9702. DOI
  • P. Sanjeewani, B. Verma, J. Affum (2021). A novel evolving classifier with a false alarm class for speed limit sign recognition. IEEE Congress on Evolutionary Computation (CEC), Kraków, Poland, pp. 2211–2217. DOI

2020

  • P. Sanjeewani, B. Verma (2020). An optimization technique for the detection of safety attributes using roadside video data. 35th IEEE International Conference on Image and Vision Computing New Zealand (IVCNZ), Wellington, New Zealand, pp. 1–6. DOI

2019

  • T.G.P. Sanjeewani, B. Verma (2019). Learning and analysis of AusRAP attributes from digital video recording for road safety. 34th IEEE International Conference on Image and Vision Computing New Zealand (IVCNZ), Dunedin, New Zealand, pp. 1–6. DOI

2016

  • T.G.P. Sanjeewani, W. M. S. M. M. Wijeyarathne, G. T. S. Dharmasena, M. A. S. T. Ireshika (2016). Designing and simulation of an energy control algorithm for photovoltaic based active generator using state flow analysis. 3rd Annual Research Symposium, Faculty of Engineering, University of Ruhuna, Galle, Sri Lanka. Link
  • T.G.P. Sanjeewani, W. M. S. M. M. Wijeyarathne, G. T. S. Dharmasena, M. A. S. T. Ireshika (2016). Microcontroller based energy controller for grid integrated DC source in a PV system. Present Around The World (PATW), organised by the IET Young Professionals' Section, Galle, Sri Lanka. Link

For the complete and most current list, please see my Google Scholar profile.

Research Projects

Funded Project · 2024

A novel automatic neural network feature extractor with explanations and non-iterative learning

Funded by Australian Research Council Discovery Project Grant, the project focused on developing an automatic neural network-based feature extraction framework with fast, non-iterative learning and explainable features. It investigated the feature extraction capabilities of convolutional and traditional neural networks, addressing challenges such as model complexity, training efficiency, and interpretability. The proposed approach aimed to support accurate and explainable feature extraction from both image and non-image data across diverse real-world applications.

With Griffith University, Australia. Supervisors: Prof. Abdul Sattar and Prof. Brijesh Verma.

Industry Project · 2023

Secure multi-platform for identifying and protecting world-heritage Aboriginal UNESCO rock art sites in Far North Queensland

Developed an AI-powered digital platform using drone imagery, machine learning, and mapping technologies to identify, document, and preserve thousands of historically significant Aboriginal rock art sites in Far North Queensland. The project involved labelling and analysing drone footage to support the creation of a high-resolution digital archive of cultural heritage sites.

With KJR, Australia, and partners (Western Yalanji Aboriginal Corporation, Athena AI, Emesent, FlyFreely, MaxusAI, World of Drones Education Pty Ltd and Griffith University), funded by Advance Queensland, Queensland Government, and Trusted Autonomous Systems (TAS). Supervisors: Dr Kelvin Ross and Dr Mark Pedersen.

Industry Project · 2022 – 2023

Automatic hard hat and safety vest detection, slip/trip/fall detection, and hooning (reckless driving) detection AI projects

Developed AI-powered video analytics solutions for real-time safety monitoring using CCTV cameras. The projects focused on detecting PPE non-compliance (hard hats and safety vests), identifying slips, trips and falls, and detecting hooning (reckless driving) through deep learning-based object detection and tracking, enabling automated alerts and incident monitoring.

With Smart AI Connect, MaxusAI (Australia) and Veracity AI (Sri Lanka). Supervisors: Christopher Lane, Glenn Neuber, John Fitzgerald and Stan Potums.

PhD Project · 2019 – 2022

A deep learning technique for the analysis of road safety attributes

Funded by the Australian Research Council (ARC) Linkage Projects, this project developed a deep learning-based automated system to analyse road videos, detect and classify road objects, and assess road safety conditions, aiming to improve road infrastructure assessment and reduce road fatalities.

Industry partners: Department of Transport and Main Roads (DTMR), Queensland, and the Australian Road Research Board (ARRB). Supervisors: Prof. Brijesh Verma and Dr Mary Tom.

Teaching & Supervision

Lecturer, Course Coordinator and Tutor

2026

  • Semester 2COSC2391 Further Programming (UG)114 students
  • Semester 1COSC2820 Advanced Programming for Data Science (PG)237 students, CES 4.14/5, OSI 4.22/5Teaching and course overall scores were above all university benchmarks in 2026
  • Semester 1COSC2815 Advanced Programming for Data Science (UG)85 students, CES 4.50/5, OSI 4.24/5Teaching overall scores and 4/5 course scores were above university benchmarks in 2026

2025

  • Semester 2COSC2820 Advanced Programming for Data Science (PG)307 students, CES 4.03/5, OSI 4.12/5Recognised among the highest OSI courses in the School of Computing Technologies in 2025 by the A/DVC
  • Semester 2COSC2815 Advanced Programming for Data Science (UG)133 students, CES 3.98/5, OSI 3.7/5

Tutor and Lab Demonstrator

2026

  • Semester 2COSC2960 Foundations of AI (UG)789 students
  • Semester 1COSC3144/COSC2527 Games and AI Techniques (PG)208 students

2025

  • Semester 2COSC2960 Foundations of AI (UG)640 students, CES 4.5/5
  • Semester 1COSC2793 Computational Machine Learning (PG)147 students
  • Semester 1COSC2673 Machine Learning (UG)289 students

2024

  • Semester 1COSC2793 Computational Machine Learning (PG)99 students, CES 4.19/5
  • Semester 1COSC2673 Machine Learning (UG)142 students
  • Semester 1, 2COSC2960 Foundations of AI for STEM (UG)943 students (semester 1), 1813 students (semester 2)
  • Semester 2COSC2670/COSC2738 Practical Data Science with Python (PG)499 students
  • Semester 2COSC2531 Programming Fundamentals (PG)535 studentsOne of the top 5 courses in the School of Computing Technologies

2023

  • Semester 2COSC2779/COSC2972 Deep Learning (PG)60 students

Earlier Teaching

Supervision

Awards & Scholarships

Honors & Awards

Top Educator recognition

RMIT University, 2025

From the Associate Deputy Vice-Chancellor (Semester 2).

Most Outstanding Paper Award

IEEE ICAC, 2025

7th IEEE International Conference on Advancements in Computing, Kotte, Sri Lanka.

Presentation Award

Central Queensland University, 2021

School of Engineering and Technology (SET) Research Higher Degree Students Symposium.

Scholarships

Professional Activities

Journal Reviewer

Conference Reviewer

Invited Talks

Conference and Symposium Presentations

Memberships

Organising Committees

Voluntary Work