•  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  

The successful candidates will lead innovative research projects in a stimulating, open, and international research environment with a highly talented and motivated team, embedded in a strong network of academic and industrial collaborators. 

The candidates will work in collaboration with 3 labs on program co-funded by the EPFL Center for Intelligent Systems (CIS):

David Atienza (STI), head of the Embedded Systems Laboratory ESL,
Adrian Ionescu (STI), head of The Nanoelectronic Devices Laboratory (NANOLAB),
Martin Jaggi (IC), head of Machine Learning and Optimization Laboratory MLO.

Main duties and responsibilities include :
Postdoc 1 – Decentralized AI & Health

The postdoc jointly with PhD students and medical collaborators contributes to advancing the practicability of decentralized machine learning and federated learning. This contrasts current approaches which require to centralize all training data and thus form a main roadblock to wider adoption in many AI applications, particularly in the medical field. We develop decentralized and federated deep learning training schemes which are efficiently scalable, privacy-preserving, robust to malicious actors, and personalized to each user’s needs. While these 4 key aspects have been partially addressed individually in the existing literature, their unification in the project here will form the main scientific contribution by the postdoc and collaborators.
On the software side, the research project will specifically aim towards interfacing deep learning models with mobile and edge AI devices. This will allow for integration with use cases personalized health applications. Produced software will be made available open source.

Postdoc 2 – Edge AI Devices for DigiPredict, Exposome and Beyond

This Postdoc will work in cooperation between two EPFL laboratories (Nanolab and ESL) as well as international medical partners. The topic to be developed in this post-doctoral position is multi-disciplinary and include the linking between NANOLAB and ESL through the development of edge AI systems from the software mapping and architecture to the inclusion of new sensors and technologies.

The postdoc will work on the development and coordination of the following works:
(i) design a multi-modal sensor interface to provide real-time data fusion from both biosensors and physiological sensors, structured for data analytics (this interface should be able to integrate also exposome data from environmental sensors at longer term),
(ii) fuse and filter in an energy-efficient way at the edge level information from multiple types of sensors, as well as on mapping federated AI algorithms for mobile and edge AI devices to enable predictive and personalized edge AI-deployed digital twins. This work includes for different tasks, namely:

– Develop a biocompatible heterogeneous system integration platform for patch integration including new specialized (multi-core) edge AI computing architectures with new innovative biosensors (e.g. based on nanotechnology or metamaterials).
– Develop new approaches to reduce the volume of information to transfer among autonomous wireless wearable sensors for real-time multi-analyte (biomarker) sensing and multiple physiological signals.
– Develop a framework to embed federated and the new decentralized AI algorithms developed within the work of the other postdoc, on the new edge platform as well as on mobile coordinator hubs for real time usage of Digital Twins in the context of the exposome application. In principle, the framework can be retargeted to other applications beyond the field of healthcare that requires multiple types of sensors.

Our target, in addition to top publications, is to produce a template of an open-source multi-core platform, as well as an open-source mapping and optimization framework for different types of federated and AI algorithms on multi-core embedded platforms.

Your profile :
Candidates should have completed, or be near completion of a PhD with a strong international publication record in areas such as (but not limited to) machine learning.

Ability and motivation to co-lead applied (interdisciplinary) projects e.g. in collaboration with medicine, industry or other sciences is a plus.

We offer :
– World-class research environment;
– Numerous collaboration opportunities with researchers from EPFL and external partners from industry and academia;
– Excellent working and living conditions.
Start date :
The starting date is flexible, but an earlier start date is preferred.

Term of employment :
Fixed-term (CDD)Duration :
1 year renewableRemark :
Only candidates who applied through EPFL website or our partner Jobup’s website will be considered.

apply online

Leave a Reply