Freelance › Freelancers services › Software development › Fortran & Scientific Computing Developer — HPC and Numerical Code
Fortran & Scientific Computing Developer — HPC and Numerical Code
Service description
I am a Fortran developer focused on scientific and high-performance computing, and I have spent years living inside numerical code that other people are afraid to touch. My daily work is writing fast, correct simulations, solvers and data-processing pipelines in modern Fortran (77 through 2008/2018), and making them run efficiently on multi-core machines and clusters. Unlike specialists who reach for Julia, MATLAB or C++ by default, I deliberately stay in the Fortran world, because for tight numerical loops, linear algebra and legacy scientific software it remains the most honest and predictable tool. I care about numerical stability, reproducibility and getting the physics right, not just the code compiling.
A large part of what I do is rescue and modernization. Research groups and engineering teams often own decades-old Fortran that still produces the results everyone depends on, but nobody remembers how it works. I refactor that code carefully, add tests and documentation, replace fixed-form spaghetti with clean modules, and wire it into build systems and CI without changing the science. When performance is the goal, I parallelize with OpenMP and MPI, tune memory access and vectorization, and profile honestly so we know where the time actually goes. I also connect Fortran cores to Python or C for tooling and visualization when that makes sense.
If you have a solver that is too slow, a simulation you need extended, or an old scientific program that must survive one more decade, I can help. I explain trade-offs in plain language, keep results verifiable at every step, and hand back code your own team can maintain after I am gone.
— Modern Fortran development (F77 through 2008/2018), modules and clean interfaces
— HPC and parallel computing with OpenMP and MPI on clusters
— Numerical methods: linear algebra, ODE/PDE solvers, integration, optimization
— Simulations and modeling for physics, engineering and applied math
— Legacy code rescue: refactoring, testing, build systems, documentation
— Performance profiling, vectorization and memory tuning
— Interfacing Fortran with Python and C for tooling and visualization
A large part of what I do is rescue and modernization. Research groups and engineering teams often own decades-old Fortran that still produces the results everyone depends on, but nobody remembers how it works. I refactor that code carefully, add tests and documentation, replace fixed-form spaghetti with clean modules, and wire it into build systems and CI without changing the science. When performance is the goal, I parallelize with OpenMP and MPI, tune memory access and vectorization, and profile honestly so we know where the time actually goes. I also connect Fortran cores to Python or C for tooling and visualization when that makes sense.
If you have a solver that is too slow, a simulation you need extended, or an old scientific program that must survive one more decade, I can help. I explain trade-offs in plain language, keep results verifiable at every step, and hand back code your own team can maintain after I am gone.
— Modern Fortran development (F77 through 2008/2018), modules and clean interfaces
— HPC and parallel computing with OpenMP and MPI on clusters
— Numerical methods: linear algebra, ODE/PDE solvers, integration, optimization
— Simulations and modeling for physics, engineering and applied math
— Legacy code rescue: refactoring, testing, build systems, documentation
— Performance profiling, vectorization and memory tuning
— Interfacing Fortran with Python and C for tooling and visualization
Contact the freelancer
Order the service or ask the freelancer a question.
Freelancer contacts
E-mailShow
