Lokos AI is led by Diego Casabuena, an entrepreneur, martial artist, and private pilot, with a rich and diverse international background. Having spent a decade in the United States, he has also lived in Venezuela, Spain, the UK, Israel, and China. He holds a double major in Computer Science and Economics from NYU, where he focused on machine learning and decision theory, complementing his studies with ad-hoc PhD-level coursework in both fields. As a student researcher at the NYU Center for Data Science, he was involved in early advancements in language model development.
Diego's professional experience spans a wide range of industries, including fintech, social media, dating, and streaming applications. At Podz, he spearheaded the development of fine-tuned language models based on podcast transcription, creating the world's first audio feed. This groundbreaking innovation led to Podz's acquisition by Spotify, where Diego served as a Senior Machine Learning Engineer.
At Spotify, Diego advanced the Podz technology to introduce podcast preview features, significantly influencing the platform's evolution. He also led the development of one of the first commercially implemented audio-visual translation systems, enabling translations for prominent podcasters like Lex Fridman. In addition, he contributed to model development in transformative projects such as Spotify DJ and Auto-Mix. Diego's publication record ranges from advanced diarization model research to large-scale engineering and economic commentary.
His expertise encompasses traditional machine learning, natural language processing (NLP), and audio, image, and video ML technologies. Beyond the realm of AI, Lokos AI brings valuable experience in robotics applications and financial modeling.
"Why use humans to train ML when you can use ML to train humans."
"You can't replace jobs that never existed in the first place."