Verified profile · /p/sample
Emily Peterman
AI Builder · Foundations
Reskilling backend engineer shipping AI products that ground answers in real data and surface uncertainty cleanly.
@formerly-acmeBrooklyn, NYBuilding on GitHub since 2017emilypeterman.dev@emilypeterman
Graded by Claude Sonnet 4.6 · rubric v1
Composite
Strong signal
Quiz
82
Project
88
Engineering — Foundations
Project artifact
Repository
sample-builder/rag-support-pilot
Narrative · excerpt
We chose hybrid retrieval (BM25 + dense) because the corpus is short, technical, and full of acronyms — pure dense missed exact-token queries by ~12 points on Recall@10. After adding a reranker, the answer-faithfulness eval moved from 71 → 87 with no latency regression at p95.
Adaptive quiz
Quiz score · 0 to 100
Correct on first try
Microskills exercised
Strongest microskills
- RAG evaluationrating 1378
- Evals designrating 1334
- Uncertainty surfacingrating 1306