Implement weighted solution ranking combining AI quality score and community votes

[Mkalbani]

Description

Solutions currently have AI-powered quality evaluation and community votes as separate signals, but there's no unified ranking endpoint that combines them. Add a documented, tunable ranking formula so solutions can be sorted by overall standing rather than by vote count or AI score in isolation.

Acceptance Criteria

• A RankingService computes a composite score from AI quality score, net votes, and recency
• Weighting coefficients are configurable via environment variables or a config file, not hardcoded in the formula
• GET /problems/:id/solutions?sort=ranked returns solutions ordered by composite score
• Ranking formula and its rationale are documented in docs/ranking.md
• Edited solutions (per the existing edit-history feature) are re-scored after edit rather than keeping a stale score
• Unit tests cover score calculation for solutions with missing AI scores or zero votes (no divide-by-zero or null-score crashes)

[Truphile]

Hi team! I'd love to take this on.
​I'll handle the core implementation, keep the state/data layers in sync, handle edge cases cleanly, and ship it with solid tests.
​Ready for assignment to drop a PR!