Migrate to gpt-oss-120b, strict JSON schema, per-IP rate limiting

Increase parse-retry budget from 1 to 4 (5 total attempts)

Retry once on EvaluationParseError to mask stochastic JSON drift

Migrate inference backend from Hugging Face to Groq

Add inference module with HF API wrapper

Add evaluation module with parsing, validation, and formatting

Initial project scaffolding