docs: add research benchmark metrics table

[CoreyLeath-code]

Summary

• Replaces the compact README metrics table with research-style benchmark and recorded metrics sections.
• Adds recorded platform targets, local smoke benchmarks, Prometheus/observability metric inventory, and project reproducibility metrics.
• Records validation limits where native Go/C++ compiler checks could not be run in the local workspace.

Validation

• pytest --basetemp=.pytest-tmp/run --cov=api --cov-report=term-missing tests/test_health.py tests/test_auth.py tests/test_inference.py tests/test_rate_limit.py
  • 4 passed
  • 24% focused api coverage
• Local synthetic benchmarks:
  • SentinelModel: 577 trainable params, 0.0160 ms avg forward-pass latency, 62,470.72 RPS synthetic throughput
  • PSI/KS Python-equivalent drift calculation: 0.001549 ms avg over 10,000 runs
  • LLM fallback summary path: 0.000104 ms avg over 10,000 runs
• git diff --check

Notes

• Installed slowapi locally to run the FastAPI test suite because it is declared in requirements.txt but was not present in the shared workspace venv.
• Native Go/C++ compile checks were not run because Go and C++ compilers were unavailable in this workspace.

Summary by CodeRabbit

• Documentation
  • Replaced the basic metrics table with a dated, comprehensive research benchmarks and recorded metrics section.
  • Added platform targets, local smoke benchmark results, observability metrics, and project reproducibility metrics.
  • Included measurement and normalization details for improved transparency and reproducibility.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 46e579bb-6846-4c63-af66-df890316afe6

📥 Commits

Reviewing files that changed from the base of the PR and between 0f8a841 and 9bc329e.

📒 Files selected for processing (1)

• README.md

---

📝 Walkthrough

Walkthrough

README.md replaces the former metrics table with dated platform benchmarks, local smoke results, observability metric definitions, and project reproducibility counts measured and normalized on 2026-07-12.

Changes

README metrics documentation

Layer / File(s)	Summary
Benchmark and reproducibility metrics README.md	Adds categorized tables for platform targets, local benchmark results, observability metrics, repository and test counts, CI/CD workflows, dependencies, delivery infrastructure, and validation limitations.

Estimated code review effort: 1 (Trivial) | ~5 minutes

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch docs/add-research-metrics-table

---

_{Comment @coderabbitai help to get the list of available commands.}