En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)
Portail > Offres > Offre UMR7334-DELSTU-020 - Conception de système mixte (mémoire RRAM et numérique/analogique) visant un accélérateur matériel de réseau neurones artificiels H/F

Design of mixed system (RRAM memory & digital/analog) targeting artificial M/W

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : mercredi 9 décembre 2020

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

Faites connaître cette offre !

General information

Reference : UMR7334-DELSTU-020
Workplace : MARSEILLE 13
Date of publication : Wednesday, November 18, 2020
Type of Contract : FTC Scientist
Contract Period : 17 months
Expected date of employment : 1 February 2021
Proportion of work : Full time
Remuneration : between 2728.25 and 3881.28 € according to experience
Desired level of education : PhD
Experience required : 1 to 4 years

Missions

Research position in the framework of the project
UNICO (Unsupervised spiking neural networks with analog memristive devices for edge
computing) supported by the Horizon 2020 FET program of the EU. The hardware of the
UNICO project will take advantage of heterogeneous integration of analog resistive
memories crossbar arrays in the Back End Of Line (BEOL) of CMOS analog/mixed
neurons. This dense, energy efficient and truly parallel architecture will implement the
synaptic connections required by SNNs while LIF neurons will be implemented in CMOS.

Activities

The candidate will contribute to design different flavors of analog/mixed neurons
compatible with the different learning rules defined at system level (e.g. Spiking Timing
Dependent Plasticity) constrained by the resistive device feature (TiO2) and will
propose behavioral model (Verilog-A) suitable for system simulation. The neurons
design will also have to integrate configurability to be compliant with different
numbers of input/output synapses and will also embed trimming functions to ease the
adaptation to various RRAM electrical regimes. The candidate will also contribute to
the test of the proposed neurons after manufacturing.

Skills

the applicant should have a
Ph.D degree in Electronics Engineering, Electrical Engineering. Solid knowledge in the
area of circuit design (preferentially analog design) is necessary. Any experience in
resistive memory and/or neural network is a plus.
English is mandatory, French is a plus but not mandatory

Work Context

The candidate will work in the IM2NP laboratory, in the Mémoires team.

Constraints and risks

No identified risks

We talk about it on Twitter!