Background & interests
About
I am a researcher and engineer interested in how AI systems work, how to make them more efficient, and how to make their behaviour easier to understand.
I hold a PhD in Computer Science from the University of Cambridge, supervised by Professor Jon Crowcroft. My work combines numerical computing, distributed learning, model inference and agent applications. I have been a core developer of Owl since 2017 and its project lead since 2020.
I work in English and Chinese, and enjoy the exchange between research, implementation, teaching and open-source maintenance.
Experience
- Oct 2023 – Oct 2026Karlsruhe Institute of Technology, Germany
Postdoctoral research in AI systems for autonomous driving. Led OverDrive, contributed to C2CBridge 2, and delivered a water-level prediction module with Covadonga.
- Feb 2021 – Apr 2023Beijing Institute of Technology, China
Postdoctoral research on distributed and federated learning, edge intelligence and industrial AI.
- 2017 – presentOwl open-source project
Core developer; project lead and release manager since 2020. This work runs alongside my research appointments.
Education
- 2015 – 2021PhD, Computer Science · University of Cambridge
Corpus Christi College. Supervisor: Professor Jon Crowcroft. China Scholarship Council scholarship.
- 2013 – 2015MEng, Software Engineering · Beijing Institute of Technology
- 2009 – 2013BSc, Software Engineering · Beijing Institute of Technology
Teaching & supervision
At KIT, I contributed to two Master's-level courses: Collective Perception in Autonomous Driving and AI for Autonomous Driving. I have co-supervised seven Master's theses across Cambridge, BIT and KIT, on computing optimisation, federated learning and autonomous-driving topics.
At Cambridge, I supervised undergraduate computer networking and volunteered through the STIMULUS programme, helping teach computing to local school pupils.
Patents & recognition
I am the first inventor on four Chinese invention patents covering probabilistic edge training, federated client selection and distributed image recognition. Six software copyright registrations arose from the distributed and federated learning work.
I was a member of the team awarded an iENA Nuremberg gold medal in 2025 for an air-ground collaborative vehicular crowdsensing system. My professional credentials include AWS Solutions Architect – Professional and NVIDIA's Generative AI LLMs certification.
Beyond work
I practise kendo and enjoy cycling. Completing the full Lake Constance cycling route is one of my favourite trips. My learning notes include course projects in language modelling, AI, graphics and GPU programming.
Contact
For research, engineering and open-source collaboration: rho.ajax@gmail.com.