The Microblink Commerce suite lets brands and retailers reimagine the way they interact with consumers, leveraging the world’s largest product catalog to bring retail items to life and to give shoppers the personalized buying experience they expect and demand. Every year, our Al-powered technology transforms more than 3 billion receipts into real-time insights that power loyalty programs, promotion strategies and market research.
This summer our Commerce Engineering team was fortunate enough to be joined by Aarushi Singh, Katherine Wang, and Santiago Garcia Santos, three Penn M & T Program students and software engineers! Between working out of our Brooklyn, NY and Zagreb, Croatia offices, we caught up with them about how they discovered Microblink, the projects they worked on, and what they think about the future of Commerce powered by Microblink.
Can you share a bit about yourself and how you discovered Microblink?
Aarushi Singh: I’m a rising junior studying Computer Science and Finance at Penn. I’m interested in Machine Learning, so getting to see how a team of ML researchers approaches a problem from start to finish—creating datasets, training models, and evaluating accuracy—has been a great experience! I discovered the software engineering internship from a post made in the M&T program’s alumni website. (Editor’s Note: Danny Panzer, VP of Engineering for Commerce, is an alum!) I loved how exciting Microblink’s applications of ML are, especially at a time when it is so hard to innovate in the space due to high saturation. All of the internship projects sounded not only interesting but also important to the company, rather than just busywork.
Katherine Wang: I’m a rising sophomore studying Computer Science and Business Analytics. I was really excited by how the company leverages ML to do meaningful work across different industries. While I was recruiting for the internship and talking to people at Microblink, I could see how dedicated everyone was to the company’s mission and how passionate they were about the products!
Santiago Garcia Santos: I’m also a rising sophomore pursuing a double major in Computer Science and Finance. Over the past few years, I’ve become pretty interested in app development and software engineering more generally, so when I saw the internship opportunities at Microblink, I knew I had to reach out. I was excited by the possibility of working with Microblink’s cutting-edge ML models and applying them in AR experiences for mobile apps.
Can you describe the projects each of you have been working on this summer?
AS: I’m working on training a model to read abbreviated receipt descriptions and translate them into full product names, including working to build the ground-truth datasets we’re using to train our models. I’ve also gotten to dive into the field of Natural Language pProcessing, specifically neural machine translation, and I’ve learned so much about the state-of-the-art model architectures in that field. I’m currently testing out different machine translation frameworks and training models in each of them to see which gives the best results!
KW: I’ve been working on an image stitching app on Android that enables users to scan large shelves or long receipts and stitches the captured frames together. In the case of shelf scanning, it will detect the products and their locations. This was my first time doing Android development, so there was so much for me to learn! I not only learned about the fundamentals of Android dev, but I also familiarized myself with computer vision principles and OpenCV (one of the biggest computer vision libraries).
SGS: We were given a lot of flexibility when choosing our projects for the summer, so I chose to capitalize on my previous experience in iOS development to help develop an iOS app able to run our product recognition ML models on device. I used this capability to create a more user-friendly, interactive experience for the data collection teams on the ground, which is crucial to continuously improving our models for better performance.
You got to spend some time with the Commerce product suite. What most excites you about what the team is working on?
AS: I’ve been working on improving the product intelligence side of our Commerce suite, so I might be biased, but what excites me most is improving our current systems to become the best in the industry! Receipt scanning and product intelligence are so critical to Microblink’s core value proposition, and it has been exciting to watch how we’re constantly improving our accuracy at every step in the pipeline – from character recognition through connecting scanned names to products.
KW: I’m most excited about the in-store discovery app that Microblink is working on and how they’re integrating ML and Augmented Reality to revolutionize the future of grocery shopping. There’s a huge opportunity to digitize parts of the grocery shopping process and create even better experiences for customers.
SGS: I’ve been lucky enough to work pretty closely with some of Microblink’s ML innovations, experiencing the speed and accuracy as I implement them in my own projects. From a technical perspective, I really think the application of this in AR has the potential to be the most exciting development for the Commerce team, truly innovating upon the in-store shopping experience. In the next few years, I think AR will start to become more and more central to how people interact with the world.
What has been the highlight of your internship experience? Did anything surprise you?
AS: Besides learning so many practical skills about how to architect and train models, the highlight of my experience has definitely been seeing how a tech startup operates from the inside. I love the weekly calls with our entire Commerce Engineering team, where I can hear all the interesting projects everyone is working on and see how they fit into the bigger picture of the company. One thing that surprised me is how fast everything moves! We can approach a problem and have a new product or model to solve it deployed in just a few weeks, all thanks to the talented and agile engineering teams.
KW: Honestly, there have been so many highlights during my time at Microblink! Everyone here is so welcoming and supportive, and it’s incredible to see all the hard work, drive, and passion required to make a company successful. I’ve really enjoyed learning in such a fast-paced work environment. I’m able to see what all the other engineers are working on, and it’s really inspiring to hear about the cool and innovative tools, apps, and projects they’re currently building. I was really surprised and impressed by how much each individual is doing for the company in terms of scope, and how they can make such a huge impact on the products Microblink has to offer.
SGS: I’ve got to say the highlight has been the incredibly supportive, welcoming and helpful team of people I’ve encountered. At the start, I was a bit apprehensive to ask people for the help I needed to get up and running, but I quickly saw just how willing everyone was to help. Something that felt very unique to the experience of working in a growing AI start-up was how open and connected the office is, including being able to chat directly to executives, engineers and more.
What has been your #1 takeaway from the internship?
AS: I think my biggest takeaway has been learning about the machine learning models in my field. I have learned so much about the model architectures, what makes them so effective, and how to implement them. The team I work with is so knowledgeable about ML, and they’ve been able to send me so many helpful research papers and answer my questions.
KW: Computer science is such a massive, rapidly-changing field, and no matter how hard anyone tries, they’ll never learn it all. I learned that it’s okay (and very normal!) to not know everything – what matters more is my excitement to learn something new every day. I definitely got better at figuring out how to approach a seemingly daunting task. During my internship, I’ve also found it super valuable to collaborate with my fellow engineers. Aarushi, Santiago, and I were all working with very different tools, but I always love hearing about what they’re working on!
SGS: Every day I’m constantly learning loads of different things, from the ins-and-outs of Xcode Project settings, to learning some C++, git best practices and more. More substantially, I’ve learned about developing in a team environment, including the importance of communication, writing code with readability in mind, as well as the whole ideation and creative process.
What’s next for you three?
AS: This internship has solidified my love of software engineering and Machine Learning. I’ve had the chance to see just how powerful computer science is and how much there is out there to learn.
KW: I’ve really enjoyed coding with the intention of serving users and creating the most enjoyable experience for them. Going forward, I hope to continue applying my computer science background to help build meaningful and exciting products!
SGS: Throughout my internship, I’ve loved when I’ve had the autonomy to ideate and develop an experience/workflow. I really like thinking about the user experience as I develop my own ideas, and hopefully I can apply this thinking as I try to develop my own apps in the future.