About me
I am a Senior Machine Learning Researcher and lead of the ML Research team at Brave. I focus on designing and building privacy-preserving, ML-based systems.
My research interests lie in the areas of Privacy-Preserving Machine Learning (Federated Learning, P2P), Mobile Systems, and Efficient On-Device ML (including LLMs). In the past, I have also been interested in Human-Computer Interaction and Robotics (Human-Robot Interaction).
I hold a Ph.D. in Computer Science from the School of Electronic Engineering and Computer Science of Queen Mary University of London. My thesis titled "Analysing Crowd Behaviours using Mobile Sensing" was focused on the application of machine learning techniques on mobile sensor data to understand how crowds behave in planned events. During my PhD I was a Co-founder and CTO of Hiver, a London Business School (LBS) startup that analysed the human interactions in social events using mobile sensing technology. I also hold an M.Sc. in Software Engineering from Queen Mary University of London.
In the past I have worked as a Researcher at Telefónica Research in Barcelona (Spain), a Research Associate in the Dyson School of Design Engineering at Imperial College London, a Research Assistant at Queen Mary University of London, and an R&D Engineer at Philips Healthcare (formerly Philips Medical Systems) in Germany.
I am also the author of SensingKit, a fairly popular open-source mobile sensing framework for iOS and Android platforms.
We are looking for talented students to join us for a 3-month summer internship. Additionally, we welcome external academic collaborators for joint research projects. If you are interested, please get in touch!
Recent News:
- New Excited to share BLaDE, our latest open-source tool and hardware infrastructure for mobile performance testing. Check out the Code and Blog post.
- New Our work "MELTing point: Mobile Evaluation of Language Transformers" has been accepted at ACM MobiCom '24, in Washington DC, USA. Check out our Paper, Code and Blog post.
- Delighted to be part of The Brave Technologist Podcast, talking about Privacy-Protecting AI and Building Consumer Trust. Tune in to learn more about our ML Research team's efforts in Privacy-Preserving Machine Learning.
- Delighted to be invited to give a talk at the Flower Summit 2022 at the University of Cambridge. Join us to hear more about our "Practical Privacy-Preserving Federated Learning" approaches at Brave Software.
- Our work "FLaaS - Practical Federated Learning as a Service for Mobile Applications" will be presented as a demo at HotMobile '22 by Nicolas Kourtellis. Preprint and code of long paper also available.
- I'm thrilled to announce that I am joining the research team of Brave as a Machine Learning Researcher.
- Our paper "BatteryLab: A Collaborative Platform for Power Monitoring" has been accepted to the Passive and Active Measurement Conference 2022 (PAM '22). Congratulations to the BatteryLab team.
- Our paper "Choosing the Best of All Worlds: Accurate, Diverse, and Novel Recommendations through Multi-Objective Reinforcement Learning" with Dušan Stamenković, Alexandros Karatzoglou, Ioannis Arapakis and Xin Xin has been accepted to the 15th International Conference on Web Search and Data Mining (WSDM '22). Congrats all!
- We are thrilled that our work "PPFL: Privacy-preserving Federated Learning with Trusted Execution Environments" has won the ACM MobiSys 2021 Best Paper Award!
- Our work "PPFL: Privacy-preserving Federated Learning with Trusted Execution Environments" with Fan (Vincent) Mo, Hamed Haddadi, Eduard Marin, Diego Perino, and Nicolas Kourtellis has been accepted to ACM MobiSys 2021 with very positive reviews. Paper, Teaser, Talk and Code are all available. Congratulations to all, but mostly Vincent for having two MobiSys papers accepted in a row ('20 and '21)!
- We are organising the 1st Workshop on Security and Privacy for Mobile AI (MAISP), co-located with ACM MobiSys 2021 in a virtual (on-line) format. Please consider submitting your work.
- Our work "FLaaS: Federated Learning as a Service" has been accepted to the 1st Workshop on Distributed Machine Learning (DistributedML 2020), co-located with CoNEXT 2020.
- Our work "DarkneTZ: Towards Model Privacy on the Edge using Trusted Execution Environments" has been accepted to ACM MobiSys 2020. Source code and video of the presentation are both available. Congratulations Vincent!!
- I'm thrilled to announce that since February 2020, I am part of the research team at Telefónica I+D in Barcelona, Spain.
- Our work "A Hybrid Deep Learning Architecture for Privacy-Preserving Mobile Analytics" has been accepted to IEEE Internet of Things (IoT) Journal, 2020.
- We are thrilled that our work "BatteryLab, A Distributed Power Monitoring Platform For Mobile Devices" has won the ACM SenSys '19 Best Demo Award.
- Vincent's ongoing work "Towards Characterizing and Limiting Information Exposure in DNN Layers" has been accepted at the 26th ACM Conference on Computer and Communications Security (CCS '19) poster session.
- Our paper "BatteryLab, A Distributed Power Monitoring Platform For Mobile Devices" has been accepted at the 18th ACM Workshop on Hot Topics in Networks (HotNets 2019), Princeton, New Jersey, November 2019. Try a live demo of BatteryLab here.
- Our work "BatteryLab, A Distributed Power Monitoring Platform For Mobile Devices" has been accepted as a demo at ACM SenSys '19, New York, November 2019.
- Our work "Finding Dory in the Crowd: Detecting Social Interactions using Multi-Modal Mobile Sensing" has been accepted at SenSys-ML '19 workshop, taking place in conjunction with ACM SenSys '19.
- Our ongoing work "High-Accuracy Energy Measurements on Mobile Devices" will be presented in this year's 2nd UK Mobile, Wearable and Ubiquitous Systems Research Symposium, July 1-2, 2019.
- Our work "Typical Phone Use Habits: Intense Use Does Not Predict Negative Well-Being" will be presented at ACM MobileHCI 2018, Sept. 3-6 in Barcelona, Spain. Read more about this work at my blog post.
- Part of our latest work "Detecting Social Interactions using Multi-Modal Mobile Sensing" will be presented in this year's 1st UK Mobile, Wearable and Ubiquitous Systems Research Symposium, Sept. 12-13, 2018.
- Part of our latest work "Detecting Social Interactions using Multi-Modal Mobile Sensing" will be presented in this year's Next Generation Networking, Multi-Service Networks workshop (Coseners 2018), July 5-6, 2018.
- Our position paper "The Potential of Wearable Tech to Monitor Social Interactions through Interpersonal Synchrony Detection" has been accepted at WearSys 2018 workshop, taking place in conjunction with ACM MobiSys 2018. Full paper is in the works.
- Part of our work “Walking in Sync: Two is Company, Three’s a Crowd” has been demonstrated in this year’s Royal Institution Christmas Lectures at BBC Four. Read more about this work at my blog post.
- Our work "Beyond interruptibility: predicting opportune moments to engage mobile phone users" will be presented at UbiComp 2017, Sept 11-15 in Maui, Hawaii. Read more about it in Martin Pielot's blog post.