Metrics SQL by Rill vs ml-intern

Metrics SQL is a SQL-based semantic layer from Rill Data that solves a specific and painful problem: AI agents that query your data warehouse tend to hallucinate aggregation logic, producing metrics that look plausible but are mathematically wrong. Metrics SQL lets analysts define business metrics once — revenue, MAU, conversion rate, ROAS — in a governed definition layer, and then exposes those definitions as queryable SQL tables. Every dashboard, notebook, and AI agent resolves from the same source. The technical approach is elegant: rather than inventing a new DSL, Metrics SQL extends SQL itself. An agent that knows SQL can query `SELECT * FROM metrics.weekly_revenue` and get correctly computed numbers without needing to know how revenue is defined, which tables it joins, or how edge cases like refunds are handled. The semantic layer intercepts the query, applies the governed definition, and returns correct results. The implications for AI-native data stacks are significant. Currently, one of the biggest failure modes for AI analysts and BI agents is inconsistent metric computation — different agents or dashboards produce different numbers for 'revenue' because they implement aggregation logic differently. Metrics SQL addresses this at the infrastructure level, not by improving agent prompting.

ml-intern is Hugging Face's own open-source autonomous ML engineering agent. Given a task description, it reads relevant papers, writes training code, executes it in a sandboxed environment, evaluates the results, iterates, and ultimately uploads a trained model to the Hugging Face Hub — with no human in the loop beyond the initial prompt. Under the hood, the agent runs an agentic loop of up to 300 iterations, using Claude as its reasoning backbone alongside smolagents. It has integrated access to HF documentation search, paper retrieval, GitHub code search, and sandboxed Python execution. When the context window fills (at 170k tokens), it auto-compacts rather than failing, and full sessions are uploaded to HF for inspection and reproducibility. What's notable here isn't just the capability — it's the source. Hugging Face is essentially shipping a proof-of-concept that the job of "write the ML training script, run it, fix it until it works, upload the result" can now be delegated to an agent. With 688 stars and active development as of this week, ml-intern is HF eating its own dog food on autonomous AI engineering. The "doom loop detector" that flags repetitive tool-use patterns is a candid acknowledgment of how agentic loops fail in practice.