더 나은 교육을 위한 데이터 파이프라인 구축

In Riiid, a lot of data flows between teams and products. The Data/MLOps team’s mission is to cleanse data, safeguard any sensitive information, and communicate with other teams. They do this by building and managing data pipelines that include data ingestion, ETL, lakehouse, data cataloging, security and governance. Let’s hear from Minyoung Kim and Bhavish Rao at the Data/MLOps team how they plan to provide support for the company’s wider mission by building the internal data hub.

Please introduce yourself. 

Minyoung My name is Minyoung Kim, and I’m the TL at the Data/MLOps team. Previously, I worked at Yahoo, Oracle, Panasonic, and

Perceptive Automata. I studied at Stanford for my master’s in computer science and I’m currently pursuing my PhD at MIT. I consider

myself a lifelong learner, and neural networks (both artificial and biological) for learning and memory are the areas of study that I’m

currently interested in.

Bhavish Hi, I’m Bhavish Rao, the Product Manager at the Data/MfL platform team. I started my journey as a developer working with 

technologies like graph databases and blockchain. I have worked on products in the sports, restaurant, and gaming industries. I like

building data-driven products and solving complex problems with the latest technologies. I’m also an adrenaline junky who enjoys

water sports and soccer.

I’ve always been passionate about building products that solve complex problems. As a product manager, my goal is to create great

experiences for users through product design and development. My favorite part of being a PM is understanding a problem that needs solving, then coming up with creative ways to solve it through technology or design.

Why did you join Riiid Labs?

Minyoung I joined Riiid Labs because I believed (and I still believe) in the company's vision of leading the education market by 

promoting equity in learning. As someone who values education and knowledge dissemination, I believe that providing education in a way that supports one's learning style and maximizes learning capability and efficiency is critical to achieving educational equity. And, I think Riiid Labs has enormous potential to develop game-changing products that are well aligned with that goal. I still remember how excited I was to join the company and be part of its journey.

Bhavish I firmly believe that the right kind of education can change the world. The standard education system needs to be disrupted, and that's what Riiid was at the center of with their innovative AI technology. I love teaching, and I'm thrilled about the concept of

personalized learning because it has the potential to make a huge impact in the Edtech industry.

I've always been interested in how technology can be used to improve education. By joining Riiid, I thought it would be a great way to make an impact on students around the world who are struggling with traditional methods of learning. The goal is to help students

learn faster and more efficiently than they ever could before by using AI.

Are there any goals you are aiming to achieve at Riiid Labs?

Minyoung I’d like to bring us one step closer to our vision by helping to scale our products out, and one area that stood out to me was data engineering. We have great products that can rapidly evolve by utilizing a centralized pipeline where data from every product can be effectively consolidated and reliably served, and building/providing such pipeline is my current goal.

Bhavish In the long term, I am interested in cultivating and driving a culture of experimentation within Riiid. The more experiments we
run, the more information we have about what works and what doesn't. I've seen companies like Netflix and Spotify scale their data
science pipelines up to run thousands of experiments per day. They're constantly testing new features and changes to their algorithms— and they're seeing results! I want Riiid to follow in their footsteps and become a company that scales up its data science pipeline so that we can make AI/ML applications that truly benefit our users.

Please introduce your team.

Minyoung Our team was formed in 2021, with only three members. We’ve been growing fast and now have a team lead, product manager, data engineers, and ML engineers. We also have our director of information security and data privacy. I'm very proud of our team members' efforts in data engineering and admire their passion and positive energy!

We try to solve problems that arise as data flows between teams and products. There could be issues like data cleansing, data privacy, communications, etc. We build and manage data pipelines that include data ingestion, ETL, lakehouse, data cataloging, security and governance, to reduce those issues. Through the pipelines, we can support teams such as data science, machine learning, applications, and business intelligence. 

We have been closely communicating with other teams and identifying their requirements in order to design the most suitable data pipelines, and recently we’ve successfully set up a pipeline for auto-tagging data of math questions. We expect to assist with other data too in the near future. On the data privacy front, we're working to get our product certified in accordance with regulations such as

GDPR.

What are you working on currently?

Bhavish We are currently working on setting up a tool that would act as a single source of truth and enable anyone within the company to search and discover data. The idea is that a user could enter a set of keywords, and the system would return all of the relevant information from all of our internal databases. This would allow us to easily track and monitor our progress on projects, as well as produce reports for any stakeholders who may be interested in our progress.

What has been the most impressive experience for you since joining the data team?

Bhavish  I think the most exciting part of this role is that it allows me to work with engineers and data practitioners from all over the company, and help them build AI features. It's a great opportunity for me to learn how different teams approach problems, and I'm excited to contribute my own ideas as well.

What’s your team’s vision and technical goals?

Minyoung  Our vision is to become an internal hub where data from varied sources are consolidated, orchestrated, and served in such a way that internal teams can better harness the power of the data, gain insights for their customers, and develop effective data-driven product strategies. Our technical objectives include integrating all the platforms we've identified as essential and serving our pipeline to our product lines.

Bhavish In the past, engineers were required to build and maintain their own infrastructure and figure out how to manage data pipelines. This was a time-consuming process that prevented them from focusing on the actual work of building products. The goal of the Data/ML team is to remove this overhead so that they can focus on building better products for our customers.

We're a team that's developing the AI backbone for all of Riiid's products. We're working on building a platform that will enable

engineers to run production-grade data pipelines and ML workflows at scale without worrying about the complexity of data

governance, privacy, security, or compliance.

We believe that the success of AI is dependent upon collaboration between PMs, Data Analysts, Data Engineers, ML Engineers, and AI researchers. We want to make it easy for you to collaborate effortlessly because we know that when you can work together

seamlessly and efficiently, you'll be able to deliver better products faster.

Ideal coworkers for your team?

Minyoung Strong global communication skills are essential because we are a diverse team and much of our work involves interacting with other teams. Being able to learn quickly and adopt new trends in data engineering is also important.

Bhavish Our team values engineers who apply user-first thinking and learn quickly. We take ownership of our individual contributions

but collaborate and communicate to ensure that everyone's work is cumulative. We are constantly experimenting and developing

prototypes during sprint cycles, so we favor those who have the ability to develop ideas from concept to operation.