Computation & Algorithms
When the problem needs a model, an algorithm, and rigor
Mathematical modeling, optimization, metaheuristics, and scientific computing for problems that demand precise solutions.
Problem
Problems I work on
Slow or incorrect algorithms
The current implementation neither reaches the right answer nor runs in reasonable time.
Missing mathematical model
The problem has not been formulated mathematically; without this, any implementation is guesswork.
Multi-criteria optimization
Decisions over complex parameters that need targeted search and sensitivity analysis.
Scientific validation
Results must be assessed with accepted statistical and scientific methods.
Approach
What I do
Mathematical modeling
Turning the problem into a precise, implementable formulation.
Algorithm design and implementation
Metaheuristics like GA, DE, and custom variants tailored to the problem.
Numerical and scientific computing
Numerical methods, simulation, and implementation with NumPy/SciPy or TypeScript equivalents.
Sensitivity analysis and validation
Assessing stability of results with respect to parameters and initial conditions.
Technical report
Documenting the model, assumptions, results, and reproducible code.
Typical deliverables
- Problem formulation and mathematical model
- Algorithm design and implementation
- Numerical solving, simulation, or optimization
- Sensitivity analysis and validation
Engagement
How engagement works
- 01
Analyze the problem
Understand structure, constraints, and success criteria.
- 02
Design the model
Mathematical formulation and choice of computational approach.
- 03
Implement and tune
Implementation, parameter tuning, and result evaluation.
- 04
Validate and deliver
Sensitivity analysis, documentation, and delivery of code and report.
Technical evidence
Related projects are available on GitHub.
- Convex QP Solver and Event-Driven Metaheuristic Optimization projects
- Research and teaching experience in modeling, optimization, numerical computing, and analysis
- Real, inspectable case studies available in Projects
FAQ
- Which tools do you use?
- Python with NumPy/SciPy and custom libraries, and TypeScript/Node.js when integration with your product is needed.
- How is this different from software development?
- The focus is on the problem core—model, algorithm, and validation—but the result can also be integrated into a usable system or API when needed.
- How are results validated?
- Through sensitivity analysis, comparison with baselines, and statistical methods when applicable.
- What is the final deliverable?
- Runnable code, model documentation, and a technical report with assumptions and limitations.
Have a problem that needs modeling and algorithms?
Describe the problem; we will talk about the right computational approach.