About

Daniel is an AI Director and systems architect with a hybrid profile spanning computational neuroscience, product engineering, and executive leadership. He designs and ships agentic AI systems in production — from single-agent copilots to multi-agent orchestration stacks with persistent memory, hybrid RAG, and cost-aware inference.

His career traces an arc from scientific research to hands-on engineering to AI leadership: a predoctoral researcher at the Instituto Cajal (CSIC) publishing in Nature Neuroscience (~17 publications, ~900 citations), then data scientist building recommenders and computer vision systems, then technical founder building the MVP that originated a company, and today directing AI programmes for a global platform.

He believes discursive clarity is part of infra: if executives cannot articulate why a model behaves, nobody will defend it inside budget committees.

Differentiators

• Science + product + leadership — rare combination of rigorous scientific method (CSIC, Nature Neuroscience), 0-to-1 product execution (built the MVP that became a company), and team leadership (up to 5 AI experts).
• Still writes code — unlike many directors, maintains active full-stack practice (TypeScript/React, Python, Kotlin) and ships personal AI-native apps end-to-end.
• Communication — oratory, science communication (+10k YouTube subscribers, national press columns), and technical writing honed over a decade.
• Deep Agentic AI expertise — not just prototyping; production multi-agent systems with memory management, evaluation harnesses, and latency/cost optimisation.

Method

The throughline across rigs, recommenders, and agent fleets is the same: design experiments you can defend, define the eval contract before the build call, and treat memory as debt rather than helpful context. Method matters more than the substrate — it survives every stack rotation, and it’s what executives actually pay for.