Dionis Lopez

I’m an AI Engineer and Computer Science PhD with 18+ years across enterprise software development and advanced machine learning. I design and deploy production-grade AI agents, RAG pipelines, and state space models (SSM/Mamba) in data-resilient environments. My focus is bridging high-level research with real-world engineering to optimize system latency, compute costs, and user conversion.

In my recent role as a Senior AI Research Engineer, I led architecture for real-time interactive AI agents and production workflows. I also developed a conversational Visual Avatar Assistant using LiveKit and specialized AI models (Ditto integration), delivering ultra-low-latency video communication.

As a Principal AI Researcher and Distributed Systems Architect, I designed and taught postgraduate Generative AI curricula and upskilled 150+ engineers across RAG systems, token optimization, and LLM fine-tuning. I pioneered structured State Space Models (SSM/Mamba) as an alternative to classic Transformers, showing Mamba reduces computational complexity from quadratic O(N2) to linear O(N) and slashes memory overhead in massive contexts. I’ve supervised and graduated 13+ bachelors and masters theses with distinction.

Across industry roles, I engineered multi-tenant SaaS and multi-agent systems where reliability and speed mattered. I designed and deployed a Lead Classification Multi-Agent System with LangGraph and LlamaIndex, reducing lead processing latency and improving campaign conversion by 22%, and I built resilient platform features for emergency response with sub-second dispatch request performance. Earlier, I delivered enterprise document ingestion solutions using ML and OCR, and I served as a Technical Project Manager for Geographic Information Systems (GIS) projects—bringing a pragmatic, engineering-first mindset to applied AI.