My teaching statement

At the core, teaching is about helping students learn and think critically about a discipline’s knowledge, recognize its central problems and design practical approaches to solve them. I taught a diverse group of students, in and out of the COVID-19 pandemic, to evaluate and build knowledge they used to solve conceptual and practical problems. This diversity was a major challenge which compelled the use of different teaching approaches and delivery modes, as I taught the Scientific Method to university undergraduates on-site, and machine learning engineering to data scientists, onsite and online. Although the methods and materials I used differed, the common thread of my teaching was to keep students focused and interested via actively learning to collaborate on a project in small groups.

As a teaching assistant in Scientific Methodology for two years, for freshmen and sophomores in Biology at Bordeaux University, I combined seminar talks, lab assignments and collaborative research projects to teach a class of 20 students to reason scientifically. While achieving that goal was predicated on learning basic knowledge in Biology, we made it a priority to interleave passive lectures and reading assignments with weekly active research in groups on a topic students chose. I created small four-person groups within which students frequently exchanged, negotiating their topic, an article’s relevance and task delegation, during weekly visits I supervised at the library. Our syllabus conveyed clear common expectations about the course’s goals, but as maturity gaps makes reaching the same goals harder for weakest students, we relied on a questionnaire to maximize shared interests and maturity levels within groups. This enabled us to orient the weak groups toward easy topics, freeing time to explain difficult concepts. We structured the course around a series of four intermediate learning outcomes and it ended with two capstone outcomes. To force students to take ownership of their learning, we first used an individual assignment, which was to take, review and organize notes during a seminar talk I gave on my PhD’s work and deliver them in writing. Each group then chose, broke down the logic and justified the importance of research articles, during in-class discussions. Understanding the logic of a paper helped them summarize an article and deliver it in a written abstract. They performed a lab experiment in Biochemistry and described it in a written lab report. The two final assignments were to compose a literature review, delivered in a written report, and to defend that review in an oral presentation illustrated with visual slides and in a Q&A session. Allowing students to choose their research topic freely motivated them to engage in class, but I also created opportunities for them to explain their research to other groups, as experts, in their own words. I sprinkled reflections, raised questions and incentivized engagement with positive feedbacks and extra credits. This helped quickly identify misconceptions and roadblocks. The course’s highlight was when I had them visit my laboratory and observe while I performed an experiment with monkeys, which made research very concrete to them.

I taught three teams of data scientists to use practical machine learning engineering skills, as a lead instructor in the banking sector, which required a different type of interactions with students, a palette of online teaching materials together with strong planning and team coordination skills. The data scientists came from different business departments, with different business cultures and missions, and had different business and engineering skills. The COVID-19 pandemic also required swift adjustment to virtualized teaching, so I designed the course so that I could teach on-site and online. The central learning goal was to be able to develop and deploy reproducible and maintainable machine learning software, which are also critical skills in research. I faced two major challenges which were new to me: keep students motivated in the virtual setting and satisfy the full-time working professionals’ tight allocated time to learn new skills, I had to plan the delivery of a palette of materials that included slide presentations, shared code repositories, open-source software, to develop a website with documentations and troubleshooting, setup an issue tracking software to ensure prompt and continuous support from the TAs that I coordinated. The goal was to enable students to prepare and review courses, while requiring minimal practice and downtime from their side. I delivered upcoming sessions’ slides, coding exercises directly embedded in the software to deploy and questions days in advance. At that career stage, students had acquired extensive collaborative skills and I channeled discussions in which they could contribute their strong expertise and in which their diverse educational backgrounds enable all to discover new thinking frameworks, approaches and skills to solve problems, fostering mutual growth. Students’ high satisfaction was attested by course rating consistently above 9/10.

Mentoring requires more personal interactions, one-to-one exchanges, strong empathetic and listening skills and a genuine interest in students’ experience and goals. I started practicing these skills when I was a university freshman as I tutored high school students in Biology. I also had the opportunity to mentor several master's students in Neurosciences and Mathematics during my PhD and during my postdoc at Stanford. As a lead data scientist, I oversaw the supervision and mentoring of junior and senior data scientists to raise their maturity levels. It is personally satisfying to know that one of my students in Scientific Methodology is today a postdoctoral researcher in paleoclimatic reconstruction and that one of my senior data scientists now supervises his own team.