AI Evals for Cross-Functional Teams — Nachiket Paranjape & Swaroop Chitlur Haridas, DoorDash
DoorDash builds its AI evaluation (evals) by allowing non-technical teams to create their own tools for testing and improving AI.
Rather than building a perfect interface from the start, the platform team created stable APIs — standardized connection points — that strategy, operations, and product teams could plug their own tools into. Evaluation became a multi-team workflow: operations staff annotate and label data, product managers set quality requirements, and engineers provide measurement data. The process itself is straightforward — they collect AI responses, mark the best examples, train a "judge" (a small AI model that grades automatically), and monitor performance over time. Since different teams have different needs, they accept that there is no single right way, and the cost per annotation job dropped significantly.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
More from AI Engineer
Inside Arca's Long-Horizon Agents That Run the Business - Aman Singh, Arca
AI Engineer 23 Oct
Real AI Agent Attacks and What Would Have Stopped Them — Andrey Rybka, Bloomberg
AI Engineer 23 Oct
Outside of tech there has been no financial payoff to US companies for using AI, now viewed as a margin-hurting cost center, as VC-backed firms burn more tokens
Techmeme 4 h ago
Source: Replit recently moved most AI spend from Anthropic to OpenAI, which gave 30%+ off in a deal worth $100M+/year, after Anthropic discontinued its discount
Techmeme 6 h ago