Rebuilds that cannot afford to fail
The system still runs the business, nobody who wrote it is here, and a big-bang rewrite is a bet you cannot cover. This is the work I am sharpest at.
The instinct with legacy is to rewrite. That instinct is almost always wrong: the old system encodes years of decisions nobody wrote down, and a rewrite discards them all at once then rediscovers them in production.
So I do not start by changing anything. I measure. On one rebuild a single page was issuing 3,982 database queries per request; it now issues 52. That number was found, not guessed, which is why the infrastructure bill fell with it.
On another — three major versions behind, lead pipeline running at volume — the upgrade went Laravel 5 to 8 in place, incrementally, without stopping.
- →Measurement first — a baseline before anything changes
- →Incremental cutover, not a big bang
- →Characterisation tests around what must not break
Something holding you back?
New build, rebuild, or adding AI to what you already have — I'll take a look and give you an honest perspective.