Weave (by Weights & Biases)

ML-focused LLM tracing, evaluation, and playground built for teams already in the W&B ecosystem.

Overview

Weave extends traditional MLOps into generative AI by providing automatic trace logging, interactive playgrounds, and programmatic evaluation decorators. It connects LLM development with existing model training workflows, emphasizing code-first instrumentation and sweep-like iteration.

Best for

• MLOps teams transitioning to LLMs
• Code-first evaluation workflows

Trade-offs

What XeroHack pre-wires

When you pick Weave (by Weights & Biases) in the interview, the engine emits these files into your scaffold:

• @weave.op()
• weave.publish()
• Weave + W&B sweep integration