Research engineer positions: Nano 2017 ESPRIT
Stream processing & pattern mining with hardware support
To apply, contact Vincent Leroy
Data production grows continuously. The development of the Internet of things and sensors produce terabits of activity traces. Pattern mining algorithms are a cornerstone of data mining. They consist in detecting frequently occurring patterns (sets, sequences, sub-graphs) In data. These patterns can later be used in many applications, including classification, recommendation and prediction. Existing approaches to pattern mining focus on batch processing, that is offline computation. However, more recent work considers stream (online) processing. Stream processing has the advantage of reading data only once, which limits the complexity at the cost of approximate results. They also allow continuous analysis, hence results can be obtained with low latency to detect anomalies in real time.
The goal of the Nano2017 ESPRIT project is to propose a hardware solution for pattern mining in high throughput data streams. This solution, which could be proposed as a support hardware card, will be able to test simultaneously the presence of a large number of patterns in the data. The benefits of such a solution are (i) to process faster streams than purely software approaches, and (ii) to use less servers to process data streams, thus reducing energy consumption.
DESIRED SKILLS AND EXPERIENCE
- A strong desire to implement systems that use the latest scientific results
- A good command of English
- Ability to work as part of a team
- Sufficient educational background to understand the science and mathematics involved in machine learning/ data mining algorithms
- Experience working in Linux/Unix environment
- Experience in C/C++ or Java
- Experience with at least one of the following: Python, Torch/Lua, Matlab
- Practical experience with machine learning, deep learning is a plus
Two-years postdoctoral position on Contextual recommendations
The goal of a recommendation strategy is to estimate a user’s interest for items she has not expressed interest for before, and return the items she is most likely to appreciate.
Context-aware recommendations refer to the need to take into account additional information in serving recommendations in serving content to users. Context refers to many different dimensions, temporal (time of day or time of year), geographical (at home or at work), presence of absence of others (in the company of friends or in the company of kids), etc. Context can be utilized at various stages of the recommendation process, including at the pre-filtering and the post-filtering stages and also as an integral part of the contextual modelling. This project aims at investigating how various techniques of using the contextual information can be combined into a single recommendation approach to improve recommendation accuracy. These techniques will be applied in a real use-case provided by our industrial partner, Total.
– Understand which dimensions are relevant for our use case
– Understand the changing nature of context
– Design algorithms for contextual recommendation
– Collaborate with marketing experts from Total to apply this research in real-world testing scenarios
DESIRED SKILLS AND EXPERIENCE
A strong desire to implement systems that use the latest scientific results
A good command of English
Ability to work as part of a team
Sufficient educational background to understand the science and mathematics involved in machine learning/ data mining algorithms
Coding proficiency in at least one of Java, C++, Python
Practical experience with recommendation systems on a variety of datasets
Give your best to better energy and make the commitment with Total. With over 500-plus professions in 130 countries, we offer high safety and environmental standards, strong ethical values, an innovation culture and wide-ranging career development. Be part of the global team whose mission is already shared by 100,000 employees: to make energy better each and every day.