Tudor Jianu
PhD Student of Artificial Intelligence in Medical Robotics
Tudor Jianu completed an MSc in Data Science and Artificial Intelligence at the University of Liverpool and is currently working towards a PhD in Artificial Intelligence focused on medical robotics. Specializes in research areas that include autonomous navigation in endovascular procedures. Beyond the academic sphere, shows a strong inclination for creating clean software solutions. Thrives in fast-paced, innovative environments committed to pushing the boundaries of medical technology.
- Location
- Liverpool, Merseyside, UK
- tudorjnu@gmail.com
- Website
- https://tudorjnu.github.io
Experience
– present
PhD Researcher at University of Liverpool
Pursuing a Ph.D. in autonomous endovascular navigation, Tudor excels in integrating deep learning, software development, and data science. Engineered an endovascular simulator using Python and MuJoCo, demonstrating expertise in areas like reinforcement learning and neural networks. Notable for problem-solving and team collaboration skills, which extend to mentoring and conducting AI workshops.
Highlights
- Developed an endovascular simulator, CathSim, using Python
- Deployed scripts on HPC's (Barkla)
- Researched and applied state-of-the-art neural network architectures
- Created an autonomous agent capable of performing endovascular navigation tasks
- Presented at the Hamlyn Symposium of Medical Robotics
- Supervised and guided trainee researchers, providing essential technical advice to foster academic development
- Developed demonstrations for SmartLab using industrial manipulators (UR5) and instance segmentation networks (Mask-RCNN)
– present
Team Member at University of Liverpool CS Outreach
The Department of Computer Science at the University of Liverpool actively engages in educational outreach programs to promote computer science. In collaboration with the Electrical Engineering and Electronics department, a range of events and resources are offered to schools and the general public.
Highlights
- Delivered an Introduction to Programming Workshop at Liverpool World Museum using Pi2Go robots
- Participated in continuous improvement by generating suggestions, engaging in problem-solving activities to support teamwork.
- Generated ideas while working in a small team to produce activities for a series of fun and engaging events to market the university
- Developed an Introduction to Machine Learning activity as part of Computer Science Summer School activities.
- Created Machine Learning activities by modifying TypeScript and HTML code
- Created various educational activities for the Computer Science Department
Volunteer
–
President of Data Science and Artificial Intelligence (DSAI) Society at University of Liverpool
Established the society
Highlights
- Organized Data Science networking and training events;
- Set aims and objectives in order to improve the society while securing the commitment of others;
- Built the committee for the society, advertised roles and advocated conflicts
–
Course Representative at University of Liverpool
Representative of Data Science and Artificial Intelligence MSc students
Highlights
- Gathered information from the students by asking open questions to better understand the views in order to improve the course
- Participated in meetings throughout the year, supported the opinions of the students and raised and complains to the department as well as offered tentative solutions to the issues.
Education
– present
PhD in Computer Science from University of Liverpool with GPA of
–
Masters of Science in Computer Science from University of Liverpool with GPA of 4.0
Courses
- COMP516 - Research Methods in Computer Science
- COMP517 - Programming Fundamentals
- COMP518 - Database and Information Systems
- COMP533 - Maths and Statistics for AI and Data Science
- COMP527 - Data Mining and Visualisation
- COMP532 - Machine Learning and BioInspired Optimisation
- COMP534 - Applied Artificial Intelligence
- COMP575 - Computational Intelligence
- COMP702 - MSc Project
–
Bachelours of Arts in Business Administration from Coventry University with GPA of 4.0
Courses
- A106MC - Design Your Own Project
- 110SAM - Internal Business Relations
- 148HRM - Managing People
- 105MKT - Marketing Essentials
- 108SAM - Quantitative Methods for Business
- 117ECN - The Economic Environment of Business
- 120SAM - Continuing Professional Development 1
- 2000ACC - Principles of Business Accounting
- 240SAM - Supply Chain and Operations Management
- A203IAE - Making Money On-Line
- 211MKT - Buyer Behaviour
- 220SAM - Continuing Professional Development 2
- 247SAM - Exploring Business Strategy
- 251FIN - Introduction to Financial Services
- 344SAM - Project Management
- 361SAM - Contemporary Business Strategy
- A320DEL - Absolute Beginners' German 3
- 320SSL - Continuing Professional Development 3
- 348SAM - Managing Change
- 352SAM - Business Dissertation
- 353FIN - International Finance
Publications
Reducing tactile sim2real domain gaps via deep texture generation networks by International Conference on Robotics and Automation (ICRA)
Engineered a neural network that synthesizes realistic textures on simulated tactile images, targeting only contact areas to enhance realism and reduce the Sim2Real accuracy gap in robotic sensing tasks.
Cathsim: An open-source simulator for autonomous cannulation by arXiv preprint arXiv:2208.01455
An open-source simulator has been introduced to advance machine learning for autonomous endovascular navigation, offering high-fidelity catheter and aorta simulation with real-time force feedback.
Unsupervised Adversarial Domain Adaptation for Sim-to-Real Transfer of Tactile Images by IEEE Transactions on Instrumentation and Measurement
ACTNet, an unsupervised adversarial network, is proposed for tactile image transfer, using correlative attention and task-related constraints to enhance sim-to-real transfer, achieving 92.85% accuracy in real-world classification without real labels.
Translating Simulation Images to X-ray Images via Multi-Scale Semantic Matching by arXiv preprint arXiv:2304.07693
A new method is proposed to convert endovascular simulator images to X-ray-like images, emphasizing structural integrity through multi-scale semantic matching, outperforming existing techniques and accompanied by a new benchmark dataset and open-source code.
Skills
- Technical Skills
- Keywords:
- Soft Skills
- Keywords:
- Programming Languages
- Keywords:
- Tools
- Keywords:
Interests
- Sports
- Keywords:
