Conversation
Highlights

Welcome to another episode of Category Visionaries, the show that explores the future of tech with innovative B2B founders. In today’s episode, we’re speaking with Jeong-Suh Choi, CEO and Founder of Bobidi, an AI training platform that has raised nearly $6 million in funding.

Key topics discussed in this episode:

• Jeong-Suh’s career journey, split between the startup world and big tech companies like eBay and Facebook (now Meta), where he learned about machine learning and the challenges of predicting and making AI models work correctly
• The inspiration behind Bobidi, which came from a holiday season business plan discussion with co-founder Sue Jiang in 2020, leading to the company’s founding six months later in June 2021
• The problem Bobidi aims to solve: helping AI companies understand how their models will perform in the real world before deployment, addressing issues like bias and unknown loopholes
• Bobidi’s solution, which connects customers’ AI models directly to their global community, allowing people to find loopholes, provide feedback, and generate high-quality, focused training data
• The company’s current focus on speech data and modeling, offering pre-packaged test data in English and Korean, as well as insights into model biases and necessary improvements
• The premise that the best people to collect data and validate AI models are those targeted by the services backed by those models, leading to Bobidi’s diverse global community of users
• The potential for users to earn $10-20 per hour as a side hustle on Bobidi, which is significant compared to minimum wages in countries like the Philippines and Kenya
• The shift from the “big data era” to a “quality data era,” and Bobidi’s unique approach of connecting customer models to their community to find loopholes and create previously unknown data universes
• The challenges of bringing innovative technology to market, and the importance of focusing on customers’ day-to-day needs and partnering with them to improve models quickly
• The excitement of working on a high-stakes problem with the potential to help the global community, as evidenced by heartfelt user stories and feedback
• The five-year vision of enabling anyone, from professors to students to data scientists, to run their own data collection and model validation challenges using Bobidi’s platform