Luo Mai

Contact
Email: luo.mai11[AT]imperial.ac.uk

About Me
I am a post-doc researcher in the Large-Scale Distributed Systems (LSDS) group of Imperial College London. I am working with Peter Pietzuch on designing next-generation machine learning systems. I obtained my PhD from Imperial College London in 2018 under the supervision of Paolo Costa and Alexander L. Wolf.

I am the recipient of a prestigious Google Fellowship in Cloud Computing. Between 2015 and 2018, I was a visiting researcher in the Cloud & Information Services Lab of Microsoft Research, Redmond.


Projects
  • FungFu: Towards Adaptive Deep Learning.
    Duration: 2018 - current
    Abstract: more details are coming later.

  • TensorLayer: A Versatile Deep Learning Library for Developers and Scientists.
    Duration: 2016 - current
    Abstract: TensorLayer is a popular open-source deep learning library based on Google TensorFlow. It provides flexible yet high-performance modules that can help the entire deep learning development workflow. It has attracted 100,000 downloads and 4000 stars on Github, and is awarded the 2017 best open-source software by ACM Multimedia Community.
    [Code repository: GitHub]

  • Flare: Adaptive Stream Processing in Large-scale Data Ingestion
    Duration: 2015 - current
    Abstract: Questioning, debugging and reasoning a big cluster through massive logs is critical for spawning data-driven decisions. We are exploring how to build a scalable analytics system for collecting, computing and delivering event summaries to numerous data stakeholders at an interactive speed. Flare is a research project during incubation. It is supported by Microsoft Research, Bing Ads and Azure.
    [More details: Chi VLDB'18]

  • NaaS: Network-as-a-Service in the Cloud
    Duration: 2012 - 2016
    Abstract: NaaS integrates current cloud computing offerings with direct yet secure access to the network infrastructure by tenants. With NaaS, tenants can easily develop and execute application-specific network services, such as custom routing, in-network data aggregation, redundancy elimination and smart caching, in order to improve communication performance. NaaS is a completed research project. It was supported by Google and UK EPSRC.
    [More details: Emu ATC'17, FLICK ATC'16, NetEx HotCloud'16, MLNet HotCloud'15, NetAgg CoNEXT'14]

Publications