About the service
I develop programs in BQN—a modern member of the APL family, where instead of cumbersome loops you use brief array notation and truly first-class functions. If your data task seems heavier than it should be, I usually express it in a few lines that read almost like the problem statement itself. My work splits into three directions. First—data transformations: reshaping, grouping, joining and cleaning tabular or nested structures, where BQN's rank-polymorphic primitives replace pages of imperative glue code. Second—numerical and algorithmic procedures: sliding windows, prefix reductions, sorting variants, combinatorics, small simulations and statistical kernels, written concisely but readably. Third—algorithm prototyping: when you need to test an idea before a whole team takes it on, I sketch it in BQN so the solution shape stays visible and cheap to change. I care about notation that a colleague can return to in six months, so I comment the intent, give worked examples, and package everything with a short README and a set of tests covering edge cases. The deliverables usually include the trains and functions themselves, a set of assertions, example inputs and outputs, and notes on complexity. I've spent many years thinking in arrays—first in old APL dialects, now mostly in BQN—and happily combine this approach with small runtimes or embedded interpreters. Describe the transformation you need and the shape of your data, and I'll return code that's small, fast enough for the job, and honest about its assumptions.
