PhD · Computer Science · Northumbria University
Designing reinforcement learning systems that make real decisions in real environments. 10+ years bridging networks, optimization, and applied AI.
My career started where systems push their limits — field deployments of LTE networks across Latin America, chasing interference in the spectrum, debugging 3GPP protocol stacks. That foundation shaped how I think: distributed, constrained, real.
Over time I moved from configuring networks to teaching them. I now design reinforcement learning algorithms for environments where decisions happen in milliseconds and the cost of a wrong choice is measurable. My PhD research at Northumbria University focused on multi-agent RL for vehicular networks and 5G edge computing, and today at Iquadrat I build production-grade RLOps pipelines that carry those ideas into hardware-in-the-loop deployment.
I'm drawn to roles where AI systems thinking, technical leadership, and delivery excellence come together — Tech Lead, Engineering Manager, or AI Project Leadership. Currently contributing to Horizon Europe and SNS JU research programmes at Iquadrat, working at the intersection of 6G network intelligence and applied AI.
Agentic AI architectures and LLM-powered systems for intelligent orchestration in distributed infrastructure. Active contributor to Horizon Europe and SNS JU projects. Open to senior IC and leadership opportunities in AI/ML engineering.
Spanning the full spectrum from theoretical RL formulations to production deployment — with a strong foundation in optimization, distributed systems, and telecommunications.
Single-agent and multi-agent RL for stochastic, resource-constrained, and dynamic environments. Policy generalization, stability, convergence analysis.
End-to-end pipelines from experiment design to production. Simulation-to-hardware validation, reproducible evaluation frameworks, and model lifecycle management.
Hands-on projects in RAG pipelines (embeddings, vector databases, retrieval), autonomous agentic systems, TF-IDF, OpenAI API integration, and LLMOps workflows.
View LLM projects on GitHub →Resource allocation, dynamic scheduling, V2X, latency-aware offloading for edge computing. Deep familiarity with 3GPP standards and network intelligence.
Stochastic modeling, Markov processes, numerical optimization. Translating mathematical formulations into validated computational models.
Led teams of 10+ engineers at Ericsson. Agile backlog ownership, Lean Six Sigma Yellow Belt, research mentorship, cross-functional stakeholder management.
Applied AI projects spanning LLM pipelines, agentic systems, and engineering tools — all publicly available on GitHub.
A practical breakdown of the full RAG pipeline — from raw text to grounded LLM responses — covering embeddings, vector databases, and context generation.
Experiments contrasting classical NLP (TF-IDF) with modern RAG using OpenAI GPT — combining embeddings and retrieval to answer queries with grounded information.
An autonomous agent that handles billing issues end-to-end — querying databases, making refund decisions, and resolving problems via API calls without human intervention.
Dual-scenario agentic system: task completion agent and email organizer with TF-IDF vectorization, RAG knowledge retrieval, and a methodology comparison notebook.
An agentic AI system powered by OpenAI's ChatGPT that generates personalised language lessons from domain-specific content — demonstrating how knowledge can be dynamically transformed into structured educational material.
Peer-reviewed contributions at the intersection of reinforcement learning, vehicular networks, and 5G resource optimization. Published at journals, and conferences.
Open to senior AI/ML engineering roles, technical leadership, and applied research collaborations. Based in Barcelona, open to remote.