In 2016, Joe Kim’s neuroscience and AI journey began at Seoul National University, where he built a Brain-Computer Interface (BCI) to control drones under Professor Han Hyuk Cho. He also contributed to SNU's research in analyzing EEG (electroencephalogram) from intellectually gifted individuals, providing insights into advanced cognitive functions.

In 2020, Joe joined Harvard University’s Neuroscience program and collaborated with Professor David Salat at Massachusetts General Hospital’s Brain Aging and Dementia Lab, where he developed a brain imaging Convolutional Neural Network (CNN), a deep learning model designed for analyzing visual data by mimicking the human visual cortex, to improve prediction of cognitive decline in Alzheimer’s disease.

In 2021, Joe’s entrepreneurial drive sparked through collaborations with SK hynix, a global semiconductor company, where he developed another Convolutional Neural Network to optimize nano-twin copper pillar microstructures for better electrical conductivity. With funding from Blackstone Inc. Joe invented a Convolutional Neural Network product placement technology, enabling user-targeted product integration in content post-filming (in use by a Big Tech).

In 2022, GPT-3.5 was released after training on text, image, and audio data. After witnessing a fundamental disconnect between traditional neural networks and the human brain’s complexity, Joe returned to his neuroscience roots to bridge that gap. At Empathic Neural Networks, Joe is building the world's first human model AI based on data and algorithms aligned with the human cognitive architecture, which he believes is the definitive path to achieving safe superintelligence.