Within the realm of hobbies, Leon engages with the piano, David’s focus revolves around darts, and Janko excels in badminton. However, their collective fervor for machine learning serves as their ultimate bond.
So, when the chance came to dive into a summer of practical ML experience, they jumped on it! 🚀
FER‘s Career Center and Microblink joined forces to create an internship opportunity that bridges theory and real-world application. Leon, David, and Janko jumped on the bandwagon and successfully secured their golden ticket for the internship. We chatted with the ambitious trio to hear about their motivations, experiences, and takeaways from being part of Microblink’s ML team.
What motivated you to apply for this summer internship in Microblink?
Leon: What drew me to apply was the company’s deep-rooted focus on ML. Microblink’s reputation for housing Croatia’s premier ML team caught my attention, and I felt an instant connection with the company’s culture and the dynamic individuals working here. The expertise each person brings to the table and their willingness to assist truly resonated with me. It felt like an environment that could nurture my potential, a sentiment reinforced during the ML intern selection process.
Janko: From everything I’d gathered, the sense I had was that the internal structure here is top-notch, with well-defined business processes, enabling the team to immerse themselves in their tasks fully. My expectations were set high for an environment that fosters substantial learning opportunities, and I’m delighted to say those expectations have been more than met. It’s made me genuinely content. I was initially looking for a technical and engineering-oriented space, but what I found here has actually exceeded those initial notions. This pleasant surprise stems from my desire to be part of a setting that inspires pushing one’s limits beyond the norm. The approach to tackling challenges is remarkably comprehensive, with a meticulous problem-solving attitude. The company’s structure seamlessly facilitates resolving issues without encountering major hindrances.
David: My goal was to dive into internships over the summer and gather real-world experience. I wanted to collaborate with professionals who’ve mastered their fields and expand my horizons beyond what classrooms offer. Microblink stood out as the prime choice for this internship stint due to having one of the most formidable Machine Learning teams in the country.
What new skills or techniques have you learned during the internship that you didn’t know before?
Janko: One of the major takeaways from this internship has been the emphasis on solid engineering practices within the realm of ML. While I had a hunch about this before, I hadn’t truly had the chance to put it into action in my previous experiences. Things like test-driven development and adhering to coding best practices were often overlooked. In my prior industry encounters, such a focus was a rarity. ML often started with the aim of just making things function without considering long-term sustainability. That’s not the approach here.
Leon: Discussing my ongoing projects with colleagues has been eye-opening. You’re presented with many viewpoints – about 20 different perspectives – and it’s a goldmine for sparking creativity. I might not have been so inclined to seek out such interactions in the past, but now, it propels me forward and significantly enhances my work quality.
David: My team’s unwavering commitment to code quality pleasantly surprised me. We’re deeply engaged in practices like mutual code reviews. I’m actively involved in these processes – for instance, taking part in pull requests on BitBucket and incorporating feedback from peers to enhance the code until it reaches its pinnacle iteratively. This approach is remarkable. Our daily team meetings also serve as a platform for candidly discussing ideas. During these interactions, concepts are tossed around, ones I might not have stumbled upon independently.
What pleasantly caught me off guard was the level of support I received. Whether it was the IT service team or my colleagues, someone was always ready to assist.
David, Machine Learning Intern at MicroblinkDuring the internship, what did you learn that will further your career in ML?
Janko: Throughout this experience, I’ve realized the immense potential of iterating on various experiments while maintaining clean and effective code. I’ve got that aspect finely tuned at the moment, and I’m just a stone’s throw away from immersing myself in experiments, which I’m genuinely looking forward to.
Leon: A significant learning for me has been gauging whether applying ML to tackle a problem is a viable approach or not. It’s about discerning if investing effort in an ML-based product at the production level is justified. And naturally, mastering the art of shaping ideas, constructing a solid Proof of Concept, designing a seamless pipeline, and rigorously evaluating the end product.
David: Now that I’ve wrapped up a few experiments, I aim to iterate upon them. I’m eager to uncover any potential flaws and identify opportunities for enhancing these experiments. This process allows me to differentiate between better and less successful experiments. The added bonus is the constructive feedback I receive from mentors and fellow team members, which is invaluable for growth.
Were there any specific challenges you encountered during your projects, and how did you overcome them?
Leon: I can’t say there were any major roadblocks. However, I did encounter a minor challenge when I hit a personal creative plateau and found it difficult to forge ahead. That’s when having a mentor proved invaluable. Their guidance, fresh perspective, and advice injected new life into my approach, propelling me forward.
Janko: In the initial stages, I ran into a bit of a snarl because I hadn’t anticipated needing to employ my software engineering expertise to this extent. I was geared up to jump straight into experimenting, but my mentor rightly steered me toward emphasizing code quality and crafting tests. Looking back, I appreciate this redirection, as it enables me to make alterations in the stability of my work confidently.
David: At the outset, I grappled with matters like permissions, equipment setup, and certain internal logistics. What pleasantly caught me off guard was the level of support I received. Whether it was the IT service team or my colleagues, someone was always ready to assist. The swift response time ensured that I never felt stuck for long.
This sense of ease is why we never feel tense; shifting into focused work mode isn’t challenging.
Janko, Machine Learning Intern at MicroblinkWhat do you like the most about the ML team dynamics?
David: The Communities of Practice have been an incredible platform for fostering knowledge exchange while fostering a sense of camaraderie. Engaging discussions are ongoing, and the atmosphere is quite vibrant. We’ve even had gatherings where we could tap into the rich reservoir of people’s experiences within the Academy. The team members are incredibly open and generous when it comes to sharing their insights.
Leon: One aspect that stood out to me is the accessibility and eagerness of mentors to provide assistance. I remember the day we received our MacBooks, and I was fumbling with the keyboard, trying to figure it out. Almost instantly, five colleagues came over to lend a hand. It’s remarkable how everyone is quick to offer support regardless of workload. Whether they pertain to ML or life in general, queries are met with a willingness to help.
Janko: What strikes me is the affable and approachable nature of everyone here. Casual conversations about everyday life seamlessly transition into discussions about work specifics and tasks at hand. This sense of ease is why we never feel tense; shifting into focused work mode isn’t challenging.
What are the key takeaways or lessons you’ll carry with you from this internship journey?
Janko: The most vital lesson I’ve gleaned centers on the pivotal role of establishing an environment within an ML team that’s primed for developing ML-centric products. This is an area where Microblink excels, setting a benchmark that few companies in Croatia have embraced. Creating such an environment isn’t a simple feat, and Microblink has executed it splendidly. As for my project, I could have achieved tenfold more progress if time were limitless. Surprisingly, I didn’t encounter an upper limit; I delved deep and committed more time than I had initially anticipated, revealing that there’s still a wealth of potential waiting to be explored.
Leon: What’s noteworthy here is that ML is approached in a manner that genuinely encapsulates its essence. It’s not relegated to a mere front-end or as an auxiliary research aspect, which I’ve observed elsewhere.
David: The methodology stands out. We were assigned new projects, untouched by anyone else. In the academic setting, we often received topics that already had a measure of research done. While some of that structure exists here, we began from scratch – starting with data and problem-solving approaches, method selection, and ongoing development. This experience has significantly enriched my skill set, and it’s a takeaway I’ll carry with me.
Having these inspired and talented interns this summer was a true joy for our Machine Learning team.
We are eagerly waiting to see what the future holds for them and wish them all the best in their (already) notable ML hands-on experience! ✨