About SaltbyPep
Built by the engineers who worked the aging floor
We are a small team of food process engineers, ML specialists, and meat scientists who spent years watching expert intuition fail to transfer across chambers, facilities, and new hires. SaltbyPep is what we built to fix that.
The problem we could not stop thinking about
Premium beef aging operates on tacit knowledge. One experienced practitioner can run a chamber program by feel, adjusting humidity and timing based on sensory cues that took years to develop. Scale that to multiple chambers, two facilities, or a new hire, and the quality variance compounds immediately.
We spent years in food process environments watching this exact failure repeat. The solution is not a better checklist or a remote thermometer. It is a real-time enzymatic process model connected to the hardware that controls the environment.
The Team
Three specialists, one obsession
We bring together the process engineering, ML systems, and food science expertise needed to build a product that works on real commercial aging operations, not just in a test lab.
Kim Young-jae
CEO & Co-Founder
Food process engineering background with years building humidity-control systems for industrial fermentation environments. Co-founding SaltbyPep to replace expert intuition with sensor data in premium beef aging.
Shin Hae-won
CTO & Co-Founder
ML systems engineering for industrial sensor networks, building real-time process control pipelines in food and chemical manufacturing over many years. Leads all sensor integration and inference work at SaltbyPep.
Lim Gi-tae
Head of Food Science
Enzyme chemistry and meat science specialist with deep expertise in collagen hydrolysis kinetics and texture endpoint measurement across multiple protein processing contexts.
Make expert-quality aging reproducible without the expert
The goal is not to replace the aging master. It is to capture what they know in a form that can run without them, scale to additional chambers, and keep producing consistent results even as staff turns over. We build for processors who want to grow a premium product without the quality cliff that comes from growing headcount.
We measure ourselves against batch-to-batch texture variance and endpoint prediction accuracy, not engagement metrics. If the data shows the model is not outperforming experienced intuition in a given operation, we do not consider the product ready for that customer.
Location
Based in Seongnam, Korea
26 Pangyo-roBundang-gu, Seongnam 13494
Gyeonggi-do, Korea
- [email protected]
- Near Pangyo Techno Valley, Bundang-gu