Yugen Omer Korat's Linkedin Analytics

I develop pipelines that use LLMs as a base component that break down otherwise unachievable tasks into manageable components, in two main areas: - Content generation: I help agencies and marketers write context-driven social content at scale by building pipelines that simulate human-like text that people enjoy. Bots with social skills that keep the conversation going, and present information in a fun & gaging way. - Information retrieval: Topic modeling, text clustering, summarization, question answering and search engines. In non-technical terms, this means extracting the specific information the user needs out of huge datasets. This is at the core of my work at MarvinLabs, as well as LifeMedic and other projects I participate in. You can check my articles here for more details about all of these projects, but I'm sure before that you have an even more burning question - ------------------------------- "Wait, what do linguists *actually* do?" Well, as it turns out, they have been writing algorithms that describe how to produce language. And they've been doing it since the days of Panini (पाणिनि) in around 500 BC. The only problem with linguistic theories is that they are often way too abstract to actually code them using bits and bytes. So as interesting and cool as it might be, nobody could quite figure out how you get all great these ideas into a computer. Until now. Don't let the word linguistics fool you - my degrees were about as programming and math heavy as they come. My studies & research involved statistical learning theory, information theory, complexity theory, deep learning, and so on. I am involved in several projects, some as a cofounder, some as an employee, others as volunteer. I provide GenAI based solutions and consultation to companies and organizations. (to linguists: don't worry, I am not going to make the claim that LLMs are any kind of models of the cognitive process involved. I am only concerned about the input-output here. And I really mostly talk about pragmatics and discourse analysis, less so on formal syntax/semantics/morphology)