KAIST StartupTing X Minds Lab Review (May)

The second StartupTing of May was joined by Minds Lab. Minds Lab is a famous AI company engaged in various businesses from algorithms to platform and AI-related services. There are many AI experts as well as students from KAIST at Minds Lab. Through the stories of relatable KAIST seniors, StartupTing with Minds Lab was a time to hear about realistic stories about the company, team, and startup experience.

How do you become an AI researcher- Brain Team Researcher Hyoungkyu Song

Researcher Hyoungkyu Song of the Brain Team is working at Minds Lab as a skilled industry personnel. Researcher Song double majored in Bio and Brain Engineering and Computer Science and participated in various school activities at KAIST. The story of researcher Song will be described in a first person point of view.

Unlike other schools, KAIST has a lot of club and activities available, so I recommend students to work on various projects and try out intern. In my case, during my first internship at Nota, I wrote work shop papers, and was able to participate in various projects in the company. For deep learning tasks, one can only grow and gain skills through working on projects. To do deep learning, there are many requirements like datasets and servers, and Minds Lab provides these resources such as GPU. In the Brain team that I belong to, there are people from various departments and we also collaborate with many departments. I am in charge of audio and graphics, and develop and tune parts that are used in services like TTS and chatbots. At the Brain team at Minds Lab, we read research papers, implement them, and upload the code to paperswithcode. We also do paper review and participate in various conferences (CVPR, ECCV, ICLR). Most importantly, I recommend interning or working at a startup because you can service the engines you develop as APIs.

From a non-major to an AI Scientist – Brain Vision Team Researcher Seung min Yang

The second lecture was by Researcher Seung min Yang of the Brain Vision Team. Researcher Yang shared his feelings while working at Minds Lab for about a year. Researcher Seung min majored in Bio and Brain Engineering, minored in Biological Sciences, and is currently at Minds Lab as a skilled industry personnel as well. The story of researcher Yang will be described in a first person point of view.

In the beginning, I lacked background knowledge of computer science. I also participated in various activities, one of which is the Cuop program, where I worked at a healthcare startup. Based on my experience, I will explain how I became a researcher, the skills AI researchers need, and what I felt about Minds Lab as a researcher. While attending seminars in various fields, I became interested in artificial intelligence, but had a lot of worries about my career because graduation was ahead. In addition, I felt my lack of basic knowledge in artificial intelligence research. After joining Minds Lab and sharing knowledge with team members, I gained the knowledge required in the field and in work. The research I conducted are related to service development.

– Skin property Classification: Skin Type, Humidity, Symptom

– Comics OCR: Optical Character Recognition to transform Webtoons into Web novels

– Clothes Segmentation &Warping (Virtual TryOn): User takes a picture of a clothing, AI removes the background and tres the virtual outfit to an avatar

The skills required as an AI researcher are basic knowledge of computer science. One must understand the infrastructure to train AI (server, GPU, etc.), algorithms, data structures (especially tensors), programming languages used in research development, and an understanding of frequently used libraries (*Python – pytorch, pillow, opencv, torchvision, etc.). It is also important to do paper review, learn SOTA Algorithm and methods, understand an edit Open Source Repository, and implement unpublished source codes. Last but not least, communication is extremely important. Within the team, it is important to exchange and acquire background knowledge based on computer/NLP/Audio and data domain divisions as well as new insights. As a researcher, this is how I felt about Minds Lab. – Infrastructure: a lot of infrastructure and resources that researchers can use, and a lot of freedom in research. – Various fields of study: high interaction with other teams, learning beyond work, chance to work in other fields – Positive work environment: flexible commuting, encouraging participation in global conferences. Thus, I recommend working at startups like Minds Lab.

How do AI Scientists work? – Brain Audio Team Researcher Wonbin Jung

The last lecturer was researcher Wonbin Jung, who joined the company most recently. Researcher Wonbin Jung is currently at the Brain Audio Team, and double majored in Computer Science and Industrial and Systems Engineering at KAIST. He is a new employee, who has been working as a skilled industry personnel since March. At the Brain Audio team, Researcher Wonbin Jung is in charge of digitalizing analog signals. Brain Audio: TTS (Twitch) / AI Mouth: Conditional TTS / Brain team: Algorithm session / Code session / General Research – To work on these, it is necessary to understand the concepts and theories of deep learning and directly experience implementation methods (Github, Python, etc.). Interest, curiosity, skills in English, and Googling become helpful assets to a researcher. Researcher Wonbin Jung shared his 3-month experience at Minds Lab in detail.

Q&A and Discussion

The Q&A session was based on pre-registered questions. The three researchers gave a more realistic and appropriate answer based on their experience at Minds Lab. They shared their experience at the company and in field and communicated with the participants. There were many questions on the study and research of the AI field, such as how non-AI-related majors can work on AI research, what AI trends beginners should follow, and what kind of research papers to read. To these questions, researchers replied that it is important to read papers from various fields, continue to improve one’s personal development, continuously study the field, and work on and enjoy various projects. Other than the pre-registered questions, there were many live questions through chat and real-time. The StartupTing will be uploaded on YouTube, so please watch if you’re interested. Hearing the stories of the three researchers, the future of Minds Lab seems very bright. AI-related services developed by Minds Lab will transcend other companies, because team members continue to work on personal development. Students who are interested and want to experience AI related services including deep learning should consider Minds Lab! We support Minds Lab J Please look forward to the next StartupTing as well. Thank you!