Data Engineering Zoomcamp: Free Data Engineering course. Register here!

DataTalks.Club

Articles Slack Events Podcast Books Courses

Career advice, learning, and featuring women in ML and AI

Season 19, episode 7 of the DataTalks.Club podcast with Isabella Bicalho

Links:

Did you like this episode? Check other episodes of the podcast, and register for new events.

Transcript

The transcripts are edited for clarity, sometimes with AI. If you notice any incorrect information, let us know.

Alexey: This week, we’re talking about continuous learning in data science. Our guest is Isabella, a machine learning engineer and data scientist with three years of hands-on AI development experience. She also has a background in computational research, contributes to open-source projects, and runs a newsletter showcasing women’s accomplishments in data science. Welcome, Isabella! (0.0)

Isabella: Thank you so much. I’m very happy to be here. (6:58)

Alexey: The questions for today’s interview were prepared by Johanna Beer. Thanks, Johanna! Isabella, let’s start with your career journey. Can you tell us about it so far? (7:01)

Career journey and shift to machine learning

Isabella: I have a somewhat unusual background. My first degree was a bachelor’s in biological science, and I always wanted to be a bioinformatician. Luckily, my university allowed me to take courses from other departments, so I enrolled in computer science classes like data structures, algorithms, and object-oriented programming. (7:15)

Isabella: After that, I did a master’s in bioinformatics, where I was introduced to machine learning. Around that time, the machine learning boom started, and I never looked back. It completely changed my career, and I realized how much I loved it. (7:15)

Alexey: So, was your shift to machine learning intentional, or was it more of a coincidence? (8:19)

Isabella: It was a mix. I always liked data and thought I’d become more of a data engineer. Statistics was my gateway into machine learning. Once I understood distributions and linear modeling, it naturally led to functions and, eventually, advanced techniques like transformers. (8:29)

Isabella: I was in the right place at the right time and fell in love with it. If I had started with a passion for machine learning, it might have felt overwhelming. (8:29)

Alexey: Where did you study? (9:12)

Isabella: I did my bachelor’s in Brazil at the University of Maranhão and my master’s in France at the University of Marseille. (9:14)

Alexey: Are you originally from Brazil? (9:29)

Isabella: Yes, I am. (9:32)

Alexey: Your last name doesn’t sound French. (9:33)

Isabella: No, I’m not French at all! But I’ve been living here for about four years. (9:37)

Alexey: Do you speak French? (9:47)

Isabella: Yes, I do. (9:48)

Alexey: Did you learn it before moving to France? (9:49)

Isabella: Yes, I learned French before moving to France. I wouldn't have dared to move without speaking the language—it’s very hard. (9:52)

Alexey: Especially if it’s not Paris. I lived in a small city in France, and it’s very difficult to get by with just English. (9:57)

Isabella: I used to live in Marseille, and it’s different because it’s such a tourist hub. You’d hear people speaking Arabic, English, and German on the streets, especially in the summer. It could get a bit crazy when cruise ships docked, and the city would fill up with tourists. I remember not being able to find water in the supermarkets because the city wasn’t equipped to handle the influx. (10:08)

Isabella: Marseille is unique in that sense, but when it comes to administrative work, you really need to speak French. (10:08)

Alexey: We also noticed on your LinkedIn profile that you’ve done several internships while studying, including one at INRIA, which is quite well known. I’m familiar with them because I started there. What did you work on at INRIA? And can you explain what INRIA is? (10:43)

Internship at INRIA and working on cancer research

Isabella: INRIA is a French research institute. Its full name translates to the National Institute for Research in Digital Science and Technology. They originally focused on automation, but that term has evolved to mean AI in their context. (11:14)

Isabella: Before starting my internship, I didn’t know much about INRIA. I learned just a week prior that they were involved in projects like Scikit-learn, which made me feel a lot of pressure—I realized I was joining a big name. (11:14)

Isabella: I got the internship through a career development module in my master’s program. They invited professionals to share their career paths, and a researcher named Sebastian gave a talk. He was a mathematician who applied his expertise to biomarkers and cancer research. I had done my bachelor’s thesis on biomarkers, so I reached out to him afterward, expressing interest in his work. He invited me to do an internship, and I worked with his team for two months. (11:14)

Isabella: We developed models to predict patient responses to immunotherapy for a specific type of lung cancer, using just a simple blood draw. It was an incredible experience because I was part of a small, collaborative team. We worked in an open space, so if I had a question while coding, I could ask someone and get immediate feedback. It was enriching and rewarding, and I’m still friends with some of the people I met there. (11:14)

Alexey: Is INRIA based in Paris? (13:30)

Isabella: No, the team I worked with had recently moved to Marseille. Originally, they were based in Bordeaux, which is on the coast. (13:31)

Alexey: I’m not sure where Bordeaux is. (13:46)

Isabella: Think of France as a hexagon. Bordeaux is on the lower left corner, near the Atlantic coast. (13:49)

Alexey: I’ve heard of Bordeaux but haven’t been there. Even though I lived in France, I only know the places I visited. Does INRIA focus exclusively on machine learning and AI? (14:00)

AI research and the role of INRIA in AI and machine learning

Isabella: No, they also conduct statistical and foundational research. However, they’ve gained a lot of recognition in AI and machine learning because they’re at the forefront of the field. For instance, they co-published research on algorithms like SPAtting or SPlaTting—I can never remember the exact name. (14:25)

Isabella: France has been heavily investing in AI innovation, and INRIA benefits from that. The government is very supportive of AI research. (14:25)

Alexey: When I think about AI in France, I think of companies like Mistral. It seems like France has become a hub for AI innovation. In contrast, there aren’t many similar companies in Germany. (14:58)

Isabella: That’s partly because President Macron has been very proactive about promoting AI. France has several prominent companies, like Mistral, Hugging Face, and startups like JustCallToo, which focuses on evaluating large language models. Paris is definitely becoming an AI hub. (15:18)

Freelancing as a machine learning engineer

Alexey: Your LinkedIn profile says you’re currently a self-employed machine learning engineer. What are you working on? (15:43)

Isabella: I’m working full-time as a freelancer. I can’t disclose the company I’m working with due to an NDA, but I’m building a recommendation system. Specifically, I’m automating the process of generating knowledge graphs for a company, which has direct applications in insurance. (15:55)

Isabella: It’s quite different from what I’ve done in the past, but I’ve been wanting to freelance full-time for a while, so I jumped at the opportunity when it came. (15:55)

Alexey: You’ve been freelancing full-time for about three months now, right? (16:34)

Isabella: Yes, but I’ve done freelancing before, mainly as a data scientist. (16:40)

Alexey: So, you’ve been freelancing for a while, but now your focus has shifted. What did you do as a data scientist before? (16:45)

Transition from research to freelancing

Isabella: Before this, I worked on a project for a company that involved aggregating and processing georeferenced open-source data. Sometimes we used machine learning, but other times it was more rule-based, using classical database merging systems. The goal was to generate maps for specific regions of interest. It was very data-focused, unlike my current role, which is more model-focused. (16:57)

Alexey: Did you start your career as a freelancer? (17:38)

Isabella: No, I began in a research-based position. I wanted to gain research experience, but I realized it wasn’t for me. (17:40)

Isabella: Research is fascinating when you see the final papers and results, but on a day-to-day basis, I found it unsatisfying. I prefer engineering, where I can work toward a clear deliverable or final product. Freelancing allows me to balance both aspects—engineering and a bit of research. (17:40)

Alexey: Did you pursue a PhD while working as a researcher? (18:40)

Isabella: I started a PhD but decided it wasn’t for me. (18:52)

Alexey: So, after deciding against a PhD, you transitioned to freelancing instead of joining a company as a full-time employee. Why? (18:57)

Isabella: I wasn’t sure what I wanted to do at first, so freelancing gave me the opportunity to explore. With freelancing, you change projects often, meet new people, and experience how different companies operate. That variety and hands-on experience were exactly what I wanted. (19:20)

Alexey: Freelancing seems more challenging because you need to find clients, and companies usually expect a certain level of experience from freelancers. Wouldn’t it have been easier to get a full-time job first? It feels like freelancing is a harder path, but you skipped the traditional route entirely. That’s quite interesting. (19:48)

Isabella: It was a bit stressful to be a freelancer. Sometimes you need to actively look for clients, find offers, and contact people. However, I don't think it's as difficult as people make it out to be. It's just a different skill set. Looking for a traditional job can also be quite challenging. (20:38)

Isabella: Eventually, I want to have a full-time job in a company, but only after gaining more hands-on experience. For now, freelancing works well for me because I enjoy working on different projects and exploring various topics. Being a freelancer allows you to focus on what you can immediately offer a company, and if you like learning new technologies, it's a great way to stay up to date. You have the flexibility to learn independently and apply new knowledge to your projects right away. (20:38)

Finding the first freelance client

Alexey: That's interesting. How did you find your first client? You were doing a PhD, realized it wasn't for you, and decided to try freelancing. What was your first step? (22:07)

Isabella: I quit my PhD when I felt it was time. I decided to either look for a job or try freelancing. I posted my CV in several places since I was already planning to move from Marseille to Lyon. (22:22)

Isabella: One day, while I was at the supermarket, I got a call from someone looking for a freelancer. They explained what they needed, and I said yes. A week later, I had my first client. It all happened very quickly. (22:22)

Alexey: That’s an interesting story. (23:07)

Isabella: I wish it were more exciting, but it’s pretty straightforward. I optimized my LinkedIn profile, posted my CV everywhere, and that was it. (23:10)

Alexey: In a way, that’s a good thing—it shows that it doesn’t have to be overly complicated, right? (23:27)

Isabella: Yes, exactly. (23:35)

Alexey: You also had a strong profile, right? (23:37)

Isabella: Yes, and I was lucky to have worked on projects outside academia. For example, I participated in an AI for Good project coordinated by Antina, which involved geospatial technologies. When the client called, I already had relevant experience because of that project. (23:39)

Isabella: The same goes for the machine learning part. I had joined a few open-source projects to gain more practical experience with machine learning. When the opportunity came, I had the skills needed. It was a mix of luck and effort on my part. (23:39)

Alexey: It sounds like more than just a little effort. Balancing an AI for Good project, learning new skills, working in academia, and pursuing a PhD is a lot. Finding the time and motivation for all of that takes dedication, doesn’t it? (24:43)

Isabella: It does, but open-source work can be so enjoyable that it doesn’t feel like effort. Many projects have incredibly passionate and supportive people. That energy keeps you motivated and excited to contribute. (25:13)

Isabella: In open source, there’s often less pressure compared to academic settings or boot camps. Deadlines and expectations are more flexible, and you can approach tasks with a sense of playfulness. You experiment, learn, and share ideas, which makes the process much more engaging and fun. (25:13)

Involvement in open-source projects and community courses

Alexey: Can you tell us more about these open-source projects? How did you get involved? (26:25)

Isabella: The first project I got deeply involved in was Hugging Face’s community course on computer vision. I stumbled upon it on their Discord server. Someone mentioned creating a channel for the project, so I joined. There was a spreadsheet for contributors, and I signed up. (26:30)

Isabella: For the next three months, I spent my evenings writing about computer vision, reviewing pull requests, and working with others. I made great friends through the course—connections I wouldn’t have otherwise. (26:30)

Isabella: The AI for Good project happened because someone I met through this course invited me. That led to more connections and other interesting projects. Once you know people with shared passions, ideas naturally flow. (26:30)

Alexey: It sounds like the key takeaway is to be active in communities and open to opportunities. You see something that resonates, jump in, and participate. (28:42)

Isabella: Exactly. It can feel scary at first, but once you take the initial step, you realize it’s manageable. (29:02)

Alexey: Especially with community projects, starting something new can be daunting. But working alongside others provides structure. You take one step at a time, and it becomes clear what to do next. (29:15)

Isabella: Absolutely. In a community, people naturally take on roles. Some focus on editing and organizing ideas, while others track deadlines. Personally, I’m not great at deadlines outside of work, so I really appreciate it when someone keeps us on schedule. (29:38)

Isabella: It’s amazing to see how different skills and approaches come together to make a project successful. (29:38)

Contributing to Hugging Face's community course on computer vision

Alexey: This community course on Hugging Face was on Discord, right? You mentioned it was about computer vision. (30:12)

Isabella: Yes. (30:20)

Alexey: So, what exactly did you do there? Was it just one of the modules, or did you have a broader role? (30:21)

Isabella: I worked mostly on the first module, which covered the basics of computer vision and image processing—how computers even "see" images, and the classical techniques used. Later on, I became a reviewer for the course. We reviewed many of the other units. The course was structured so that once everyone finished their work, they made a pull request. Then, there were a lot of pull requests to review, so we would step in when available. I also contributed to sections like "Introduction to Transfer Learning." By the time I joined, most of the advanced topics were already taken, and the introductory topics were still open. So, I worked on one of those. (30:27)

Alexey: How much experience in computer vision did you have at that time? (31:25)

Isabella: When I was working in academia, both my last internship during my master's degree and the beginning of my academic career involved computer vision. So, I was already quite familiar with it. (31:29)

Alexey: I'm curious—how deep should your knowledge be to participate in a community course like that, so you don't feel like an imposter? (31:43)

Isabella: That's a great question. I was very comfortable with the topic I worked on, but there were definitely people with less experience than me. They did a great job, though. Sometimes they would ask, "I don't understand this. Can you recommend a paper or a book to help?" Even if they weren't experts in the field, they put in the effort to research. And since there was always someone more knowledgeable, we were able to guide each other and review each other's work. There were also people reviewing the pull requests, which helped catch mistakes or clarify things. So, while not all topics are beginner-friendly, I believe there's always a part of a project that is. (31:57)

Alexey: Right. And when you're experienced in a topic like computer vision, some things might seem obvious to you, and you might not take the time to explain them. But someone with a fresh perspective might find it confusing and ask how you transitioned from one concept to another. (32:55)

Isabella: Exactly. That's one of the things I love about teaching. When you work on something every day, you forget that what you find intuitive might not be obvious to others. When students ask questions, it makes you realize that not everyone has the same level of knowledge. It forces you to slow down and explain things more clearly. Also, machine learning has a lot of jargon, and sometimes terms are used interchangeably when they don't need to be. For someone starting out, it's especially important to break things down simply and clearly. (33:27)

Finding open-source project opportunities

Alexey: Let's say I want to find a project to contribute to—perhaps an open-source project. Where would you suggest looking for these opportunities? Could it be random communities, or are there specific places where people can get involved? (34:11)

Isabella: Well, first, if you use a specific open-source model or library, they likely have a Slack or Discord community. A good place to start is by joining those and seeing what others are working on. You can also check out the issues on GitHub. Often, there are easy issues to fix. If you're using a library regularly, you may spot something that's not optimal, and you could open an issue or propose a fix. Just be polite and open about it, and things will work out. Another good place is documentation. Sometimes, projects explicitly ask for contributions to documentation, which is a great place to start because everyone wants to contribute to the code, but no one wants to write documentation. Starting there can give you a good experience. On the community side, I think Omten A is a good place to look for projects. They organize local chapters where you can get hands-on experience with a machine learning project in a low-pressure environment. Training is usually provided in these courses, and everyone works together to review a couple of notebooks. I think those are two good places to begin. (34:41)

Alexey: When I think of contributing to open-source, I usually think of tools like Scikit-learn, Pandas, or NumPy. When someone says, "You can contribute to open-source," I think, "Okay, I'll go to the GitHub issues for Scikit-learn and fix something." But now, after hearing you talk about contributing to open-source projects, I realize there are different kinds of projects—it's not just about working on tools, but using those tools to solve other problems. (36:44)

Isabella: Exactly. Open-source has many dimensions. There are open-source codebases, open-source data, and open-source tools. Contributing to open-source tools is great, but it's also one of the most overwhelming places to start. If you look at a big library like NumPy and check out their issues, you’ll often see that someone else is already working on it. That doesn't mean you can't contribute, but the smaller libraries are often more welcoming to new contributors. Open-source data is also overlooked, but it’s valuable. If you can contribute to a project with open-source code, data, and a final product, that’s something really special. (37:26)

Alexey: You mentioned the community course on Hugging Face. I assume it was also an open-source project, which is very close to what we do at DataTalks.Club. We use GitHub for our courses, so everyone can contribute. While it's not open-source in the same way as a tool like Pandas, it’s still open for learning and contribution. (38:23)

Isabella: Yes, it's great. I’ve participated in some boot camps, mostly for specific models I was interested in. I really appreciate it because finding good, reliable sources of information has become more difficult. If you search something on Google, the first pages are often filled with entries that look like they were generated by AI, and they’re not very helpful. But when you find content made by people who truly understand the topic and put in the effort, it really stands out. (38:54)

Alexey: Right. And when it comes to projects, we talked about courses. You mentioned that one of the places you found your project was through Omten A, right? (39:31)

Isabella: Yes, that’s where I found my project. Actually, one of the people I met during the first project invited me to organize another chapter with her. In December, we’ll be organizing a chapter on Green Space Segmentation in Frankfurt. If anyone is interested, they’re welcome to join us. I’d be happy to welcome them! (39:44)

Green space segmentation

Alexey: Green space segmentation? (40:08)

Isabella: Green space segmentation. (40:10)

Alexey: What exactly is that? (40:12)

Isabella: It’s essentially identifying areas like gardens or forests in the city. For example, using satellite images, we can identify green spaces. The most commonly used satellite for this is Sentinel-2, which orbits Earth and provides free images. There are also high-quality drone images available. The challenge is processing these images to create a segmentation map of the city—identifying where the gardens, forests, and other green spaces are. The goal is to map this back to ground truth data. Right now, we're benchmarking different strategies, from CNNs to transformer-based approaches, to see which one works best. It’s a bit ambitious, but we want to find a solution that’s both effective and practical. (40:12)

Isabella: The issue is that sometimes research models perform well in theory, but in real life, they’re not as usable because they require a lot of computational resources. For example, if predicting one image takes an entire day, it becomes unfeasible for most people. Our aim is to get the green space map out there, but we also want to find a way that others can replicate it in other cities efficiently. (40:12)

Alexey: It sounds similar to a project you worked on before, which also involved geospatial data, right? When you were looking for your first freelance job, the company you ended up working for needed the exact skills you developed during that project, correct? (42:01)

Isabella: Yes, exactly! That’s how I landed the job, and it might happen again with this project. I think it's a great story—learning hands-on skills through projects. With platforms like Dina, there’s no expectation of working long hours. You can contribute a few hours a week, maybe after work or during the weekend, and collaborate with others to solve the problem. It’s a great way to learn new skills without feeling overwhelmed, especially when you work with a team. (42:24)

Alexey: Would you say that the skills you acquire through such hands-on projects translate directly into job skills? (43:07)

Isabella: I think so. Everyone gives the same advice in data science: build your portfolio, gain experience, and so on. (43:17)

Alexey: There are many ways to do that, right? (43:27)

Isabella: Exactly. There are many ways to build your portfolio. But as I’ve grown in my career, I’ve realized that employers also look for soft skills. They don’t just want someone who’s great at their job but can’t work well with others. When you work on these projects, you naturally pick up soft skills. If 80 people are working on the same project, you need to communicate effectively, stay organized, and manage your time. It's a great way to learn how to navigate situations like receiving a lot of messages at once and figuring out how to prioritize tasks. These are critical skills in engineering and data science. The mindset you bring to the job is just as important as your technical skills. (43:28)

Alexey: When you work alone, you don’t always know which skills are most relevant. But on a project, there are people with experience who can guide you, right? (44:31)

Isabella: Yes, exactly. You get to work with people who have more technical expertise, and in applied projects, experts in the field often join to help. For instance, in a previous project on grass segmentation, we worked with people who specialized in grass monitoring. They were eager to help because it also benefited their work. It’s a great environment because you're exposed to people with different types of expertise. (44:47)

Collaborating with experts in open-source projects

Alexey: They probably take more of a consultant role, guiding you on the best practices, while you focus on the implementation? (45:24)

Isabella: Exactly. And I always say, if you’re interested in a job, the best way to get insights is to reach out to someone with that job title on LinkedIn and ask them questions. It might be intimidating, but worst case, they don’t reply. Best case, you get some great advice. In projects like Dina, communication is more informal, so it’s easier to ask, “What do you actually do in your job? What does a typical day look like?” This kind of direct interaction is valuable because many people entering data science don’t fully understand the different roles, and job titles can be confusing. (45:34)

Alexey: When you say "company X," do you mean an actual company? (46:39)

Isabella: No, I meant an imaginary company X! (46:44)

Alexey: Got it! You mentioned 80 people—was it really 80? (46:49)

Isabella: Yes, 80 people. (46:53)

Alexey: Wow, that’s a lot. It must have been overwhelming! (46:55)

Isabella: It starts with 80, but over time, people drop out. By the end, we had about 25 people still actively working on the project. It’s chaotic at first, but things settle down. (47:07)

Alexey: What do people need to do to start working on a project like this? Is the entry requirement low? (47:24)

Isabella: Yes, the bar is pretty low. You just need to be willing to learn and sign a code of conduct to ensure you behave appropriately. That’s it. (47:36)

Alexey: So, some people realize it’s not for them or they don’t have enough time, and eventually, the number of active participants drops from 80 to 25? (47:54)

Isabella: Exactly. Almost all open-source projects have a high turnover rate. It's just part of the nature of these projects. (48:08)

Alexey: You have to plan for that when organizing such a project. (48:22)

Isabella: Yes, exactly. But it’s a good thing. The people who remain are genuinely interested, and those who leave often return when they have a better understanding of what’s involved. It’s okay to step away if your circumstances change. There's no penalty for realizing you can’t commit, but you still gain valuable experience from the project. Realistic expectations are key. (48:29)

Alexey: So, it’s not like you join, leave after a week, and then Dina is closed to you forever. If, after six months, you realize what you want to do and have more time and a clearer idea of what’s involved, you can rejoin? (49:29)

Isabella: Exactly. The door is always open. And what I’ve learned is that sometimes people want to contribute but don’t know the terms. For example, a complete beginner might not know what "push to the branch" means. You have to be clear and patient. Some people need more guidance, and that's fine. I once worked with a 17-year-old student who didn’t know where to start, so I asked him to look up five models on the Hugging Face hub and check their performance. He didn’t mind doing this because he had the time, and in the process, he learned a lot. It’s a great way to collaborate and learn from each other. (49:55)

Using tools like ChatGPT in learning

Alexey: Now, with tools like ChatGPT, you can simply tell them what to discuss and potentially learn from it. (51:28)

Isabella: Yes, that's true. But in my experience, especially since I also teach AI, I’ve found that when people go too deep into ChatGPT, they get less out of it. For example, if you ask it a very advanced question, like how to validate a segmentation model, ChatGPT will give you an advanced answer even if you ask it to simplify. It's about vocabulary and understanding—the answer might be there, but you might not fully understand it. (51:34)

Alexey: I experience this problem myself. When I’m working on things I’m familiar with, the answers usually make sense because I already have the foundational knowledge. But for things I’m new to, like Arduino, which involves hardware, ChatGPT sometimes gives me advanced answers that I can’t apply. I end up needing to break things down step by step and ask a lot of questions. After having ten separate chats, it gets overwhelming. (52:09)

Isabella: I completely understand. That happens to me too. For example, I’m on Linux, and I had an issue with my audio. I asked ChatGPT how to fix it, but it gave me terrible advice. I thought, “Okay, this isn’t helpful.” I had to go look for forums or better sources. (52:43)

Alexey: And sometimes you come across even more confusing advice. At least that’s been my experience. (53:13)

Isabella: Yes, it can be like that, but it’s still interesting. I really enjoy using it. (53:20)

Alexey: Life is never boring when you’re working with Linux. (53:24)

Isabella: No, definitely not. I had some friends in college, and we used to joke that when you know nothing about Linux, you’re very responsible. But once you know a lot, you become reckless. It's when you know just enough to feel confident but lack the knowledge to tell if you’re doing something right or wrong. That’s when you end up doing weird things. (53:29)

Alexey: That makes sense. Now, let’s talk about your Substack. Tell us more about it—what do you focus on and work on there? (53:59)

Women in data and the Data Like Substack

Isabella: I have a Substack called Data Like, which I co-created with Denny. We focus on women in data, which is a broad field—everything from data analysis and visualization to machine learning. It’s something I’ve been reflecting on for a while, especially after working with Denny in the past. Around Christmas last year, I thought about doing something similar again, and I started researching women in data. I realized the best way to contribute was by sharing stories of women working in the field. So, on the Substack, we interview women in data science and machine learning, asking how they got started, what they’ve learned, what they like about their jobs, and what advice they have. We’ve been doing this for about 10 months, since launching in January. (54:06)

Alexey: That sounds really interesting. It reminds me of an interview we did a year ago with Sara from Morocco. She had a similar initiative—an all-women WhatsApp group. They participated in hackathons one after another and often won. You should definitely connect with her. (56:03)

Isabella: That’s super cool. I think the goal of our Substack is to make these stories more accessible. When you do group work, it can get messy after a while, and people change their schedules. So, I wanted to do something that would have a broader impact. It’s also a way for me to give back to the community and amplify more women’s voices. (56:57)

Alexey: Can you share some details about your most recent Substack issues? What have you been covering lately? (57:19)

Substack interviews with women in data science

Isabella: Our latest issue came out today, featuring Bianca Haberman, a researcher in bioinformatics. She shared how she got started and discussed her work on a research tool she developed. It took her 10 years to develop this tool, and she faced a lot of skepticism at the beginning. She had to prove that the tool worked and that you could use it for automated data analysis. Over time, people reached out to her to learn more and even invited her to workshops. Our second-to-last issue covered a woman working in fake news detection, a project she developed as part of a Google fellowship. We've featured a variety of profiles, including someone who works in AI ethics and sustainability. She spoke about how AI was once seen as experimental, lacking mathematical rigor. Back then, people didn’t take AI seriously, but that’s certainly changed now. One of the most touching stories was about a woman who created a YouTube channel to teach machine learning in Aruba in the local language. It was a unique initiative and had a real-life impact. (57:27)

Alexey: Where is Aruba located? Is it in Africa? (59:59)

Empowering stories of women in data

Isabella: No, it’s actually in the Caribbean, near Venezuela. (1:00:02)

Alexey: Oh, I see. I’ve heard of the language there but didn’t know where it was located. (1:00:03)

Isabella: Yes, it’s between Venezuela and the Netherlands. (1:00:08)

Alexey: So, each issue of your Substack is an interview with a woman who’s achieved something in the data field, and you aim to share their story. (1:00:13)

Isabella: Exactly. It’s really cool because some people have typical backgrounds, while others did a total career 180. Each person offers valuable advice, and you can always take something useful from their story. One of the most memorable interviews was with someone who started as a LinkedIn recruiter. She shared how she got started—she simply filled out a form, and they contacted her. It sounds simple, but it’s a valuable lesson: If you want something, just ask for it. (1:00:22)

Alexey: It really shows how things are simpler than they appear sometimes. (1:01:11)

Isabella: Yes, and it’s quite empowering too. It’s motivating because now you know that the only thing standing between you and becoming a course instructor is filling out a form. (1:01:15)

Alexey: Exactly. It forces you to put yourself out there. One thing I regret from the start of my career is overthinking everything. It’s something you learn from, and at worst, you get rejected. But rejection isn’t the end of the world. The key is to keep putting yourself out there. (1:01:26)

Alexey: If someone listening wants to share their story with you, what’s the best way to contact you? (1:02:06)

Isabella: You can reach me via LinkedIn, or if you’re on Substack, you can send us a message through our Data Like page. (1:02:16)

Alexey: Got it. I think I found it. There’s a link that says “Data Like” and describes the newsletter as spotlighting women’s journeys in the data field. (1:02:28)

Isabella: Yes, that’s the one. (1:02:48)

Alexey: Great! I’ll share the link in the chat. I think that’s all we have time for today. Thank you for sharing your story—it was inspiring and will surely motivate many people. (1:02:49)

Isabella: I hope so. Thanks for having me. If anyone wants to reach out, feel free to message me on LinkedIn or Twitter. I might take a little while to reply, but I always respond eventually. (1:03:23)

Alexey: Thank you, Isabella. Thanks, everyone, for joining us today. Have a great weekend, and we’ll see you soon. (1:03:42)

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.

DataTalks.Club. Hosted on GitHub Pages. We use cookies.