At a Glance
- Tasks: Design and assess scheduling strategies for fault-tolerant algorithms in a cutting-edge research environment.
- Company: Join the ROMA team at ENS Lyon, part of the prestigious LIP laboratory.
- Benefits: Opportunity to continue into a funded PhD thesis with expert supervision.
- Other info: Dynamic research setting with excellent mentorship and career development opportunities.
- Why this job: Make a real impact on high-performance computing and tackle challenging error detection problems.
- Qualifications: Knowledge in algorithm design, complexity, and probabilities is essential.
The predicted salary is between 22500 - 27500 £ per year.
Anne Benoit and Yves Robert are looking for their next M2 student.
This internship is expected to continue with a PhD thesis, for which funding from the NumPEx PEPR program is already secured. The successful candidate will join the ROMA team, at the ENS Lyon in the LIP laboratory.
Duration: 6 months
Starting date: Spring, 2025 (with a PhD thesis starting Fall 2025).
The Internship (and PhD thesis) will be supervised by Anne Benoit and Yves Robert. Co-advising will also come from Emmanuel Agullo.
Context: Launched in 2023 for a duration of 6 years, The NumPEx PEPR aims to contribute to the design and development of numerical methods and software components that will equip future European Exascale and post-Exascale machines. NumPEx also aims to support scientific and industrial applications in fully exploiting their potentials.
Several error sources may impact the execution of iterative algorithms on large-scale platforms. They include fail-stop errors, that are immediately detected, and silent errors (e.g., silent data corruptions), that can be detected through some verification mechanism. Fail-stop errors correspond to permanent failures, e.g., processor crashes. Silent errors are disruptions that strike and stay undetected until they manifest eventually through strange application behaviour. Silent errors arise from two main sources: computation errors and memory bit-flips.
Protecting algorithms and software libraries from all these errors is a major concern within the HPC community. The standard way to deal with fail-stop errors is checkpoint-restart, and the optimal checkpointing period is well known, at least for memoryless IID error inter-arrival times. However, mitigating the impact of silent errors remains an open challenge. On the one hand, replication (or even triplication to avoid a sequential re-execution) does a perfect job but at a prohibitive cost. On the other hand, numerous application-specific detectors have been introduced, such as Algorithm-Based Fault Tolerance (ABFT) checksums, recomputing a residual, checking orthogonality of some vectors, applying space and time filters across a neighbourhood, etc. These detectors are usually limited to a particular error type. A major problem is that they may well either fail to detect some errors, or raise many false alarms. In other words, these detectors are not perfect: their recall and precision are not at 100%. Most, if not all published works assume perfect detectors, which is not realistic.
Mission: The first (and main) objective of this internship is to design and assess scheduling strategies based upon a combination of checkpoints and imperfect detectors to guarantee protection from a single source of silent errors with a high probability. This requires introducing some assumptions, such as upper bounding the latency of the detection, or introducing randomized tests on the data. The second step (that may come later during a PhD) is to provide a resilient holistic methodology to protect iterative algorithms from all error types, namely fail-stop errors and all sources of silent errors.
Required Skills: Some knowledge in algorithm design, complexity, and probabilities. The work is on the algorithmic side of the problem, with potential simulations to validate the results.
For further information, please contact Anne Benoit, Yves Robert, or Emmanuel Agullo.
Internship on Fault-tolerant scheduling strategies for iterative algorithms employer: Programme de recherche NumPEx
ENS Lyon offers a dynamic and collaborative work environment within the LIP laboratory, where interns are not only supported in their research but also encouraged to grow through mentorship from leading experts in high-performance computing. The internship provides a unique opportunity to engage in cutting-edge projects that contribute to significant advancements in the field, all while being part of a vibrant academic community that values innovation and knowledge sharing.
Contact Details:
Programme de recherche NumPEx Recruitment Team
StudySmarter Expert Advice🤫
We think this is how you could land Internship on Fault-tolerant scheduling strategies for iterative algorithms
✨Join Data-Science Meetups
Get yourself along to local data-science meetups or workshops. They're goldmines for networking, and you'll learn from industry pros who might just point you in the direction of internships. Plus, discussing the latest trends with like-minded individuals can really amp up your game.
✨Utilise University Career Services
Check in with your uni's career services since they often have connections with companies looking for interns. They might even organise information sessions with firms, which can be a great chance for you to learn more about potential internships and make some key contacts.
✨Show Off Your Stuff on GitHub
If you're into data science, having a GitHub profile with your projects is essential. Make sure your portfolio is public and showcases your best work! Recruiters love to see your coding skills and problem-solving approach, and it’s a brilliant way to stand out.
✨Apply Directly on Our Website
Don’t forget to check out the internships listed on our site! It's always a good idea to apply directly through our website because it makes your application easier for our team to find, and you might just catch the hiring manager’s eye by showcasing exactly what you're passionate about in data science.
We think you need these skills to ace Internship on Fault-tolerant scheduling strategies for iterative algorithms
Some tips for your application 🫡
Show Off Your Technical Skills:For a data science internship, we want to see those analytical skills shine! List your programming languages, like Python or R, and make sure to highlight any relevant projects or courses you've completed. If you've dabbled with tools like Pandas, NumPy, or machine learning algorithms, don’t hold back – include those in your CV!
Share Your Curiosity in Your Cover Letter:As an intern, your motivation and eagerness to learn are key! In your cover letter, talk about specific data science concepts that excite you and how this internship at Programme de recherche NumPEx will help you grow. Share what you hope to achieve and how you plan to tackle real-world data problems - we love enthusiasm!
Include Any Relevant Certifications:If you've earned any certifications, such as from Coursera or DataCamp, make sure to include these in your application. They show us that you're proactive and committed to expanding your data science skillset. This could make a real difference in how we assess your application!
Keep It Relevant and Concise:Remember, as an intern, you don’t need to have decades of experience. Focus on showcasing relevant coursework, personal projects, or even related volunteer work in data science. Keep your CV and cover letter concise but impactful – we appreciate clear and straightforward communication!
How to prepare for a job interview at Programme de recherche NumPEx
✨Brush Up on Your Coding Skills
As a data science intern, you might get grilled on your programming skills. Expect to tackle some coding challenges using languages like Python or R. We recommend practising basic algorithms or data manipulation tasks so you can show off your tech skills with confidence.
✨Show Off Your Projects
Prepare to discuss any projects you’ve done, whether in your studies or on your own time. Having a strong portfolio of data analyses or machine learning models will really set you apart. We can use platforms like GitHub to showcase your work to impress Programme de recherche NumPEx.
✨Know Your Stats and ML Basics
Brush up on your statistics and machine learning concepts because interviewers love to dig into this! Be ready to explain your understanding of algorithms or how you would approach a given data problem. This will highlight your theoretical background alongside your practical skills.
✨Be Eager to Learn and Adapt
Internships are all about potential and growth. Make sure you convey your eagerness to learn and adapt to new tools or methodologies. Show Programme de recherche NumPEx that you’re not just looking for experience, but that you're keen to contribute and grow within the team.
