[Apologies for Cross-Posting]
The Data Science Department at EURECOM, Sophia-Antipolis France, invites
applications for a Research Engineer position. The candidate will work
as part of a department-wide project, which aims at developing our
research activities on large language models (LLMs).
We are looking for fresh, ambitious and hard-working software engineers
with a clear passion for turning ideas in the field of machine learning
into fully-functional prototypes, demonstrators, and systems, using
cutting edge machine learning libraries, and computing facilities.
The position is funded by a French National project on Artificial
Intelligence, and targets the development of LLMs in several flavors.
Our first goal is to target an audience of EURECOM staff, including
researchers and administration, to offer novel conversational agents
that can use internal, multimodal information to interact with users for
Q&A, summarization, in-filling and many other tasks. A second ambitious
goal is to offer innovative educational services to EURECOM students, by
designing novel conversational agents instructed to follow Socratic
interactions. Such conversational agents will be augmented with teaching
material (lecture notes, slides, video lectures, code from labs, etc…)
which should be used and referenced in conversations.
The successful candidate will work in close relation with professors,
postdoctoral researchers and Ph.D. candidates, and will lead
contributions to: (i) the implementation and deployment of state-of-the
art, open-source LLMs, including inner components (novel attention
mechanisms, novel tokenization schemes, key-value caching, to name a
few) and outer components (such as retrieval augmented generation) (ii)
the definition of appropriate methodologies for training and, most
importantly, fine-tuning of our internal LLMs, (iii) the definition of
appropriate benchmarks and validation methodologies, and (iv)
communication to a public of developers, engineers and practitioners
through technical blogs, online demonstrators (e.g. HuggingFace spaces,
on-site deployment) of our results, as well as to demonstrate the
various releases of the prototype to EURECOM management teams.
Since our activities are fueled by an important public fund, in the
context of the French Government plan for AI, called the 3IA, the
successful candidate will also participate to project meetings and
reviews, as well as public demonstrations of research results.
Requirements:
* Education Level / Degree: Ms.c. or Ph.D. degree in Computer Science,
Applied Mathematics, Physics, or a closely related area with a strong
background on algorithmic development and software engineering
* Applicants should have a good background on machine learning, a taste
for learning new topics related to machine learning research, as well as
proven experience in the design of solid software engineering artifacts
* Applicants should have strong communication skills, with target
audiences ranging from the scientific and academic community, as well as
engineers and product managers
* The working language in the department is English.
Application
The application must include:
* Detailed curriculum, including (if available) a list of software
engineering projects, public demonstrators, and contributions to
open-source projects
* a cover letter describing the applicant’s interests
the contact details of 2/3 persons that can provide references about the
candidate
* the transcripts of courses taken at graduate (and optionally
undergraduate) level
Applications should be submitted by e-mail to
pietro.michia...@eurecom.fr and secretar...@eurecom.fr with the
reference : DS/PM/EFELIA-LLM/102023
Applications will be accepted until the position is filled.
Start date: ASAP
--
Raphaël Troncy
EURECOM, Campus SophiaTech
Data Science Department
450 route des Chappes, 06410 Biot, France.
e-mail: raphael.tro...@eurecom.fr & raphael.tro...@gmail.com
Tel: +33 (0)4 - 9300 8242
Fax: +33 (0)4 - 9000 8200
Web: http://www.eurecom.fr/~troncy/
_______________________________________________
Corpora mailing list -- corpora@list.elra.info
https://list.elra.info/mailman3/postorius/lists/corpora.list.elra.info/
To unsubscribe send an email to corpora-le...@list.elra.info
