Summary
Mykyta Storozhenko is an AI Engineer with nine years of experience building production-ready, multimodal LLM systems that prioritize low latency, reliability, and cost efficiency. He designs end-to-end stacks—retrieval/RAG, agents, streaming execution, and observability—using tools like vLLM, Langfuse, pgvector, Pinecone and FastAPI to turn SOTA models into practical features. At startups he scaled ingestion and LLM pipelines to tens of thousands of structured items, cut p95 latency ~3× via parallel retrieval and streaming, and achieved 2–10× lower unit costs through eval-driven model routing. He blends a strong research background (PhD studies in AI philosophy) with hands-on product engineering, shipping vision-enabled multimodal agents and CV+LLM asset pipelines. Based in the U.S. with experience across cloud providers and edge tooling, he focuses on making search that actually finds, agents that finish, and UX that feels instant. An unconventional detail: he’s shipped profitable consumer AI apps (including an LLM-driven AI-Tarot and recursive self-prompting essay generator) alongside enterprise systems.
9 years of coding experience
5 years of employment as a software developer
PhD Studies AI Philosophy LLMs Computation, PhD Studies AI Philosophy LLMs Computation at University of Kentucky
Master's degree Logic Computation Metaphysics Analytic Philosophy, Master's degree Logic Computation Metaphysics Analytic Philosophy at Kent State University
Bachelor's degree Philosophy Logic Analytic Philosophy, Bachelor's degree Philosophy Logic Analytic Philosophy at Florida Atlantic University