About the service
I develop high-performance functional software in the Clean language and have been doing it long enough to understand where its strengths truly pay off. My specialization is uniqueness typing: I model state so that the compiler can prove a value is used exactly once. This allows the Clean language to mutate arrays, records, and files in-place, while the source code remains purely functional and referentially transparent. This is how I get C-level performance from pure, side-effect-free code. Typical tasks: porting numeric or symbolic kernels to Clean, eliminating subtle reference-aliasing bugs in an existing module, designing an I/O layer via *World so effects remain explicit and ordered, and tuning natively compiled binaries to a state where hot loops truly vectorize. I profile honestly, annotate strictness where laziness is expensive, and redesign unique data structures to reduce allocations rather than just silence the type checker. Research work is also welcome: in Clean I've experimented with graph reduction, parser combinators, and reproducible benchmarks. You get readable modules, a test suite, measurements against your baseline version, and a short document explaining the uniqueness attributes so your team can develop the code without me. If you're unsure whether Clean is right for your task — send a description, and I'll tell you honestly upfront, before you spend budget.
