Simulation & AI for Scientific Research
Finite and virtual element models, PINNs, hybrid ML-PDE approaches
Predictive Modeling with Sparse or Experimental Data
Custom inference pipelines using Bayesian optimization and surrogate models
RAG AI & NLP Solutions for Technical Domains
Retrieval-augmented pipelines for patent analysis, scientific search, and internal knowledge bases
Process Optimization & Monitoring
Simulation-backed automation and MLOps for manufacturing, energy, or logistics
Technical Advisory & Prototyping
Strategy, model evaluation, or R&D MVPs for deeptech, fintech, and lab-based environments
Sebastia Naranjo
AI specialist and scientific computing researcher with 10+ years’ experience in modeling and inference for complex, data-scarce problems. Ph.D. in Applied Math; worked with Los Alamos, UNAL, Milan-Bicocca, and AI startups.
Finite and Virtual Element Methods for PDEs in magneto
hydrodynamics
Custom inference models for semiconductor manufacturing
Data analysis and behavior modeling in e-commerce
AI for science using techniques like PINNs, ray-tune optimization, and RAG systems
Finite and Virtual Element Methods for PDEs in magneto
hydrodynamics
Custom inference models for semiconductor manufacturing
Data analysis and behavior modeling in e-commerce
AI for science using techniques like PINNs, ray-tune optimization, and RAG systems
Pythonis my primary language (numpy, scipy, pytorch, ray-tune, DSPy, OpenAI API), and I also work withC++andCUDAwhen performance is critical.
Scientific publications:
AI specialist and scientific computing researcher with 10+ years’ experience in modeling and inference for complex, data-scarce problems. Ph.D. in Applied Math; worked with Los Alamos, UNAL, Milan-Bicocca, and AI startups.
Article:
Click to more info
Y. Bang, L. Medina, M. Da Silva, S. Naranjo, M. Chopra , A method for achieving sub-
2nm across-wafer uniformity performance. In Advanced Etch Technology and Process Integration for
Nanopatterning XII (Vol. 12499, pp. 35-44). SPIE.
Article:
Click to more info
S. Naranjo Alvarez, L. Beirao da Veiga, F. Dassi, V. A. Bokil, V. Gyrya, G. Manzini,
The Virtual Element Method for a 2D incompressible MHD system. Mathematics and Computers in
Simulation, ScienceDirect, vol. 211, 2023.
Article:
Click to more info
S. Naranjo Alvarez, V. A. Bokil, V. Gyrya, Manzini G., A virtual element method for the
coupled system of magnetohydrodynamics. book chapter in “The Virtual Element Method and its
Applications”, Springer, 2021.
Article:
Click to more info
S. Naranjo Alvarez, V. A. Bokil, V. Gyrya, Manzini G., A virtual element method for
resistive magnetohydrodynamics.. Computer Methods in Applied Mechanics and Engineering, Elsevier.
vol. 381, 2021.
Alejandro Tabares
Senior data scientist and systems thinker with a physics background, specializing in risk modeling and MLOps. Led high-impact projects in finance and infrastructure, including automating $13B in credit risk at Bancolombia.
My background includes:
Monte Carlo simulations and multi-state models for credit risk
Longitudinal and survival analysis for time-sensitive systems
Data integration across diverse formats (PDF, HTML, JSON)
Design and deployment of smart city infrastructure via IoT data pipelines
Education and outreach in academic and civic tech environments