Mikel
AI Reserch Scientist
Mikel Val
My passion for artificial intelligence drives my work, focusing on groundbreaking projects that blend creative thinking with cutting-edge technology.
About Me
My passion for artificial intelligence drives my work, focusing on groundbreaking projects that blend creative thinking with cutting-edge technology.
About MeHello! I'm Mikel Val Calvo, a Senior Researcher and AI Expert at the HUMAN-TECH Research Center, Universidad Politécnica de Valencia. I lead the MindFrameAI research line, integrating LLMs, Virtual Reality, and robotics to develop cognitive architectures and empathetic agents.
With a PhD in AI and extensive experience in machine learning, computer vision, and NLP, I've led multidisciplinary projects in forensic intelligence, digital health, organizational neuroscience, and neuroprosthetics. I'm passionate about creating AI-driven solutions that bridge cutting-edge technology with real-world impact.
I serve as organizer of the international IWINAC/ICINAC congress, a prestigious biennial conference on Natural and Artificial Computation, and have received the UPV Excellence Award for Knowledge Transfer 2025 for the NeuroCash project with Banco de España.
AI researcher and technical leader with 10+ years of experience in developing intelligent systems across healthcare, robotics, security, and human-computer interaction.
LLMs, RAG, ReAct Agents, OpenAI, Hugging Face, LangChain, LangGraph
PyTorch, Scikit-Learn, OpenCV, Facial Coding, Eye-tracking
EEG Analysis, Brain-Computer Interfaces, Neurofeedback, Emotion Recognition, Biosignal Processing
Project Management, Cross-functional Teams, Technical Coordination
Deep dives into AI systems, neural interfaces, and emotion recognition technologies
Multi-threaded PyQt5 system for synchronised acquisition of BVP, GSR, Temperature, and Accelerometer data. Features TCP/IP trigger server, robust HRV peak detection with O(n) artifact removal, GSR tonic/phasic decomposition, and EDF+ export. DOI: 10.5281/zenodo.3759262
Semi-automatic knowledge graph generation leveraging GPT-4, RAG, and multi-agent architectures. Features self-repairing mechanisms for RML mapping correction, Schema.org alignment, and FSM-based workflow orchestration. Achieves 95% time reduction with 97% mapping validity.
Automated spike sorting pipeline combining UMAP dimensionality reduction, Louvain clustering, and template matching. Features phase-space DTW validation, 115 spike templates, and PyQt5 GUI for dynamic script management. Robust discrimination of neural units with multi-criteria validation.
Deep dives into NeuraViPeR algorithms, EEG artifact removal, and emotion recognition pipelines
Valencia, Spain
HUMAN-TECH Research Center
Universidad Politécnica de Valencia