HPC and radioastronomy algorithms specialist

About this position

Contract : Short term research contract (CDD), full time, 1 year

Level :

PhD with work experience (up to 4 years).

Remuneration :

Gross salary of between €3,002 and €3,493 permonth, depending on experience.

Workplace :

Observatoire de la Côte d’Azur – 96 Boulevard de l'Observatoire, 06300 Nice, France

https://www.oca.eu

The specialist will work within the “Signal and Image” team in the J.-L. Lagrange laboratory which is part of Observatoire de la Côte d’Azur (OCA), on the Valrose campus of Université Côte d’Azur. He or she will interact with members of several different teams within the Franco-Swiss ANR/FNS SEAMS project and will occasionally travel in France and Switzerland.

Work context

OCA is part of the network of Observatories for Universe Sciences (OSU), a national organisation which structures the French academic research in the field of Earth and Universe sciences under the supervision of the ministry of higher education, research and space (MESRE). The Lagrange laboratory within it conducts theoretical, instrumental and observational research to study the Universe and contributes to developing and exploiting large research infrastructures (TGIR) on the ground and in space according to a national roadmap defined by the MESRE and conducted by CNRS-INSU in a context of intense international collaboration and competition.

Astronomy is a discipline now facing big data challenges (complexity, volume) as a result of observation. This calls for efficient ways to exploit this data. This offer relates to the SEAMS project which targets the development of a chip accelerating radio-astronomy workflows and its use in a High Performance Computing (HPC) context, with possible application to the SKA project.

OCA is involved in SKA through its hosting the direction of SKA France, its contributions to the development of the Science Data Porcessors and the ECLAT laboratory for upstream R&D.

Missions

The specialist will contribute to demonstrating the outcomes of the SEAMS project which is carried out within ECLAT in collaboration with Observatoire de Paris, EFPL and INSA Rennes. This contribution will happen at the interface between designing radio-astronomy software and using the PREESM tool for mapping and scheduling. Its purpose is to explore different approaches related to parallelising a radio-astronomy imaging workflow, resource usage and performance based on models of the workflow and of the target infrastructure.

This workflow faces scaling challenges for instruments like the SKA telescopes when the volume of interferometric data and the size of the images are multiplied manyfold. The primary objective is to explore the design space relative to parallelisation along the time, frequency and baseline axes in the Fourier space. This exploration will require a fine description of the interferometric data to allow for adjusting the parallelisation to the required processing.

This could then be expanded by considering how parallelisation is carried out in the image space based on facets. Validating PREESM’s estimates will require executing the workflow for a small number of hardware (accelerators) and execution configurations. This will rely on using a combination of code included in the software model and code generated by PREESM as a result of the selected mapping and scheduling scenario.

Activities

The specialist will work on developing radio-astronomy imaging code and the corresponding PREESM model. He or she will rely on existing codes and models to this end and work with the developers of the PREESM tool to use it and provide feedback on the features developed as part of the SEAMS project to support accelerators. He or she will conduct numerical experiments to collect data and compare measurements with PREESM’s predictions for validation purposes.

This work requires a fine understanding of the radio-astronomy imaging workflow as well as expertise in HPC in order to design and describe the workflow in a way which allows mapping and scheduling to correspond to the interferometric data so as to achieve performance and scaling benefits.

The specialist will carry out the following activities:

Exploit existing radio-astronomy workflows to design one allowing for jointly parallelising along the afore-mentioned axes,

contribute new methods to this workflow,

build on the existing PREESM models to model this workflow in order to carry out design space exploration and code generation,

use PREESM and provide feedback to the tool’s developers,

run the workflow on the test platform, which includes GPU and FPGA accelerators,

collaborate with scientific teams, to validate the results in terms of quality with radio-astronomers, scaling and performance with computer scientists,

report on the development, write scientific papers and present at conferences.

Skills :

Expertise on image reconstruction for radio-astronomy and experience with the corresponding workflows

Signal and image processing : applied mathematics, inverse problems in imaging

Experience with HPC environments : hardware architecture, parallelising, scheduling

Software development: Python, C/C++, Cuda, OpenCL, MPI, OpenMP, scripts, collaborative development techniques (devops, git, CI/CD)

Experience in working on Linux platforms

Scientific communication skills

Proficiency in English (B2 level written and spoken) for writing documents and attending meetings

Ability to collaborate with researchers and other engineers involved for team work

Creative and innovative mindset

Experience with PREESM is desirable

Application deadline

20 November 2026 (to take up the post on 1 January 2027)

Contacts

Please send your resume and cover letter by email to

recrutement@oca.eu

. Applicants will first be selected based on the information provided before being interviewed.

All our positions are open to people with disabilities. Our recruiting is based on skill, without discrimination as to origin, age or gender.