# LLM evaluation

Nicholas Hopper has used LLM evaluation in his work at APX Stream and Memector.

- **Category:** Models
- **Also called:** Evals
- **Used at:** [APX Stream](https://nicholashopper.com/work/apx-stream.md), [Memector](https://nicholashopper.com/projects/memector.md)
- **Updated:** November 17, 2026

## Where Nicholas Hopper has used LLM evaluation

- Built the evaluation suite that gates every prompt and model change on scored regression sets. ([AI Engineer, APX Stream](https://nicholashopper.com/work/apx-stream.md))
- Wrote the specifications and review standards for prompts, evals and agent code, and advises leadership on build versus buy, model choice and cost. ([AI Engineer, APX Stream](https://nicholashopper.com/work/apx-stream.md))
- Added an LLM-as-judge entailment check on drafted answers that fails closed when the judge is unavailable or its output cannot be parsed. ([AI Engineer, APX Stream](https://nicholashopper.com/work/apx-stream.md))
- Built golden-set benches for document extraction tools, scored on recall, precision and exact-answer accuracy. ([AI Engineer, APX Stream](https://nicholashopper.com/work/apx-stream.md))
- Measured recall@1 of 94% and recall@5 of 100% on 50 vague queries that do not contain the template's name, with median latency around 100 ms. ([Memector](https://nicholashopper.com/projects/memector.md))

---

Canonical page: https://nicholashopper.com/skills/evals  
Last updated: 2026-11-17  
Index of all pages: https://nicholashopper.com/llms.txt
