Olivier is a Machine Learning Specialist Solutions Architect based in Lyon, France, he joined AWS in 2018 having previously worked in the Amazon Marketplace team since 2014.
HereAtAWS:Tell us about your career journey. What brought you to your current role at AWS?
Olivier: I first heard about AWS at school in 2012, and the platform has been attractive to me since then because of its breakthrough features for computing and data analysis. However, my entry door to Amazon was not AWS but the Amazon Marketplace e-commerce team in France. After a little less than two years practicing data analytics and discovering applied machine learning, I applied to an overseas position at Amazon Studio in Santa Monica, California. The relocation to the US occurred four months after my wedding so it felt a bit like a really long honeymoon!
We really enjoyed discovering a new country and new cultures. We also enjoyed the proximity between the office and the Sierra Nevada. At Amazon Studio, I was able to further develop my machine learning skills and my AWS knowledge. Then, in early 2018, I saw a Paris-based AWS Machine Learning Specialist Solution Architect (ML SA) position on our internal job board, which immediately caught my attention. It depicted a pleasant mixture of machine learning, systems architecture and customer-facing problem solving. Exactly what I’ve been looking for! After going through internal interviews, I secured the new position and relocated back to France, a little less than three years after my first Amazon relocation from Paris to Los Angeles. In addition to starting a new and exciting job, I was pleased to move back to my home country.
HereAtAWS: Can you explain your role in broad terms; who are your customers and how do you help them?
Olivier: As a Machine Learning Solutions Architect my mission consists of three objectives: (1) helping our customers build machine learning systems (2) growing the awareness about machine learning possibilities and associated AWS services (3) increasing our organisational capabilities by training my peers and closing the feedback loop between customers and AWS product teams.
Our customers expect me to help them navigate the AWS offering, finding the best tools for their use-case and designing end-to-end systems that are secured, scalable, performant and cost-effective. In practice this often consists in whiteboarding sessions, product demonstrations, technical presentations and deep-dive self-paced research.
HereAtAWS: What skills do you need to succeed in your role?
Olivier: Empathy is important. Being able to take the customer's perspective helps us make suggestions truly useful for them. Also, humility and curiosity are valued in the ML domain. The research is flourishing and production deployments ramp up quickly. Consequently, best practices and tooling keep being developed, updated and improved. It is a permanent discovery and learning effort. Technically speaking, there are two important domains: (1) the ML science and (2) the ML systems engineering. Because my team do not develop many custom models (a task usually done by scientists at AWS with our customers’), the mathematics baseline is reasonably low: algebra, derivatives and trigonometry are usually enough to understand algorithms and customer challenges. On the systems engineering side, it is important to have baseline knowledge in networking, information security, web serving and data management systems. Those pre-requisites are reasonably straightforward to acquire with a mix of theory and practice.
HereAtAWS: Have you had to learn any specific new skills (technical or soft) for your role?
Olivier: When I joined the role my ML experience was mostly with structured data and for analytics purpose (causal inference, forecasting). Typically, I didn’t have much experience with deep learning nor with computer vision. Similarly, I was not experienced with Amazon SageMaker, an AWS ML platform that was released about six months before I applied to the job. I was lucky to have a three-month “ramp-up time” to learn and prepare for the new job. I took advantage of that time to learn and practice a lot, and when I started meeting customers, I was comfortable with both deep learning and Amazon SageMaker.
HereAtAWS: How does your work with customers help to make a positive impact on society?
Olivier: In the past 2 years, my team was lucky to interact with many customers from several industries that have a positive impact on society. This includes:
- Digital platforms that improve education by personalising learning paths to student skills and needs
- Agriculture customers that wanted to use machine learning to make farming more respectful of animals
- Anti-fraud systems that protect end-users online
HereAtAWS: What’s the most exciting part of your job?
Olivier: The most exciting part of the job is the diversity of customers and peers. Our customers are incredibly diverse, from a use-case, organisation, geography and requirements perspective. This diversity is very rewarding because the learning is permanent. It is truly amazing, and it is what wakes me up every morning. I interacted with customers on topics as diverse as embedded ML, distributed training, anomaly detection, graph representation, recommender systems, object detection, open-set classification, time series forecasting, and many more. My peers also come from incredibly diverse backgrounds and this is much appreciated: complementarity brings good learnings and support in every possible situation.
HereAtAWS:What’s the most challenging part of your job?
Olivier: The ML community is frantic with new ideas and new tools and the amount of knowledge and content generated every day can be bewildering! I found one effective way to filter signal from noise is to keep looking at things from the customer perspective and always start with a set of questions and priorities. This drastically filters necessary learnings.
HereAtAWS:Do you have any needs or commitment that require flexibility in your role?
Olivier: In my first year at AWS I travelled regularly. To keep the rhythm healthy and maintain family time, I was able to work from home when needed. In my past six years at Amazon I have had numerous managers, and all of them have been caring and super respectful of personal needs.
HereAtAWS: What do you love to do outside work?
Olivier: I like to spend quality time with my family and friends. Also, I’m passionate about animals and wilderness, and love to read about nature. When I’m not thinking about ML, I’m dreaming about wilderness trips in Alaska, Africa or Australia. As I just moved to Lyon, I’m also looking forward to spending time with the family in the Alps.
HereAtAWS:What advice would you give people joining AWS?
Olivier: First, people should never hesitate to apply to jobs to fit their interest! New-joiners to AWS should take advantage of their first few months to meet peers, learn about their customers and prioritize their learnings. Once in the field, it is important to seek for peer and customer feedback. Receiving feedback is the best way to build confidence and refrain any “impostor syndrome” that may arise in the first few months (everyone gets it!).
HereAtAWS:What three words would you use to describe workdays at AWS?
Olivier: Learnings, excitement, python!