docs(README): add 'Is this for me?' section (who/why/when)

[pinodeca]

Improves the README to help devs landing on GitHub answer "does this apply to me?" within ~30 seconds.

Adds an Is this for me? section after the Quick Example covering:

• The core idea — durable graph of SQL steps with per-step checkpointing
• Pain points it addresses — long-running SQL falling over, workflow logic leaking into app code, no native fan-out/branch/loop, scripts with no memory
• Who it's for — backend engineers, data/platform engineers, DBAs/SREs, AI pipeline teams
• Workloads it fits — ETL, scheduled jobs, parallel aggregation, approvals, DB maintenance, HTTP-from-SQL
• What you're probably doing today instead — pg_cron + jobs table + worker, external orchestrators, queues + workers, plpgsql with EXCEPTION blocks
• How it's different — lives inside Postgres, SQL is the workflow language, per-step checkpointing with replay, Postgres as source of truth
• What changes in your architecture — delete bespoke queues/state tables, move background work into Postgres, observability via SELECT
• When NOT to use it — sub-ms hot paths, massive stream processing, no PG17/extensions, pure single-statement SQL, heterogeneous distributed sagas

Kept scannable (bullets, no long prose) so it stays within the ~30-second skim budget. All existing sections (Features, Quick Example, How It Works, Installation, Multi-User Setup, CI, Testing, Architecture, etc.) are unchanged.