Mathematical Engineer in Scientific Computing & Modeling 
-
🔬 Currently: Quantitative Analyst at CEIE, UTE University, developing advanced data analytics workflows, predictive models, and automated educational assessment systems.
-
📊 Expertise: Physics-Informed Neural Networks (PINNs), statistical modeling, psychometric analysis, and business intelligence solutions with Power BI.
-
🚀 Recent Focus: Student dropout prediction models (88% accuracy), automated evaluation systems, and innovative educational analytics using ensemble methods (LightGBM, XGBoost).
-
🧠 Research: Active research in PINNs for solving differential equations, with published work on biharmonic equations with discontinuous nonlinearities.
-
⚡ Passion: Bridging mathematical theory with practical AI/ML applications, from algorithmic trading to educational technology.
-
📫 Reach me:
- 🏫 Educational Analytics: Developing comprehensive dashboards for formative, summative, and diagnostic evaluations with psychometric analysis integration
- 🤖 Predictive Modeling: Building ML models for student success prediction using survival analysis and ensemble methods
- ⚙️ Automation: Creating end-to-end automated workflows for data processing, report generation, and performance tracking
- 📈 Business Intelligence: Designing Power BI solutions with optimized DAX formulas and automated ETL processes
Advanced: Python (NumPy, Pandas, SciPy, Matplotlib, Seaborn), MATLAB, SQL/PostgreSQL, Power BI, LaTeX
ML/DL: TensorFlow, PyTorch, scikit-learn, LightGBM, XGBoost, Pyro (Probabilistic Programming)
Database: PostgreSQL, MongoDB, Neo4j, Vector Databases
Web: Django, Flask
Specialized: H3 (Geospatial), Physics-Informed Neural Networks (PINNs)
A biharmonic equation with discontinuous nonlinearities. Eduardo Arias, Marco Calahorrano, Alfonso Castro.
Electronic Journal of Differential Equations, Vol. 2024, No. 15, pp. 1-9, 2024.
📖 View article
Applied dual variational principle to prove existence of non-trivial solutions for biharmonic equations with discontinuous nonlinearities.
🎯 Student Success Analytics
- Developed ensemble ML model (LightGBM + XGBoost) achieving 88% accuracy in dropout prediction
- Implemented survival analysis for academic risk identification
- Integrated socioeconomic, competency, and psychological assessment data
🧮 Physics-Informed Neural Networks (PINNs)
- Research collaboration on transparent ML for differential equations
- GPU-optimized architecture with domain decomposition
- Solving direct/inverse PDE problems with mathematical rigor
📈 Educational Assessment Automation
- Automated psychometric analysis pipeline using IRT/CTT
- Power BI dashboards with real-time KPI tracking
- Reduced manual reporting time by 90%
💹 Algorithmic Trading Platform
- LSTM-based market risk analysis achieving 33% loss reduction
- Real-time data processing pipeline handling 10K+ daily transactions
- Implemented advanced feature engineering for financial time series
🌍 Geospatial Risk Modeling
- Insurability index using H3 hexagonal binning and Bayesian modeling
- Interactive risk visualization with geographic anomaly detection
- REST API backend serving 100+ concurrent spatial queries
- Physics-Informed Machine Learning: Bridging traditional numerical methods with deep learning
- Educational Data Science: Predictive modeling for academic success and institutional effectiveness
- Stochastic Processes: Applications in finance, risk assessment, and decision-making
- Variational Calculus & PDEs: Mathematical modeling for real-world phenomena
- Bayesian Methods: Uncertainty quantification and probabilistic modeling
- 10+ automated workflows deployed in production, saving 200+ hours monthly
- 3 major predictive models implemented with measurable business impact
- 1 peer-reviewed publication in international mathematics journal
"Converting complex problems into measurable, effective solutions through strategic data utilization."
📧 Contact: mat.eduardo.arias@outlook.com
🔗 LinkedIn: eduardo-arias-3e0