Method
Prompt optimization
Automatic prompt search and tuning (DSPy, GEPA, hill-climbing).
8 cases from 7 companies apply Prompt optimization. 5 report results in production.
Typical tasks: Evaluate models, Answer questions over data, Label data with LLMs, Build an assistant over internal docs and 2 more. Companies: DoorDash, Pinterest, Yandex, Avito, Netflix and 2 more. Often combined with: LLM-as-a-judge, Agent evaluation, LLM labeling, Evaluation regression gate, Golden set / benchmark and 19 more.
| How Pinterest Improved Analytics Agent Compliance with Loop Engineering | social | 2026-11 | |
| Building Food Metadata with LLM Juries, Context Optimization & Multimodal AI | food-delivery | 2026-07 | |
| Building Ask DoorDash (Part 3): Evaluation | food-delivery | 2026-06 | |
The 52x Multiplier: How Zepto Mastered AI Agent Evaluation at Scale![]() | ![]() | e-commerce | 2026-06 |
| User-Guided LLM Answer Quality Evaluation | search-engine | 2026-05 | |
When RAG Isn't Enough: How and Why We Deployed an LLM Agent in Support![]() | ![]() | general | 2026-05 |
| Deploying LLMs in Moderation: From PoC to the Target Solution | classifieds | 2026-05 | |
| Evaluating Netflix Show Synopses with LLM-as-a-Judge | media-streaming | 2026-04 |

