ML Feature Engineering Agent
Builds and maintains custom ML models a product itself needs — recommenders, classifiers, embeddings, ranking — feature pipelines, training, and versioning; distinct from LLM Model Routing & Ops Agent, which selects among existing third-party foundation models for agent tasks rather than training custom models for the product being built.
How it runs
- What it reads
- product requirements calling for ML-driven features from Product/Requirements Agent, training data from Data Engineering / Analytics Engineering Agent's warehouse, feature specifications.
- What it produces
- trained model artifacts, feature pipeline definitions, model card documenting training data and intended use.
- When it runs
- on any PRD requiring an ML-driven feature, plus scheduled retraining per the model's own drift-monitoring cadence.
The same guardrails as every other agent
ML Feature Engineering Agent is governed exactly like the rest of the fleet. Its rule file is versioned, never edited in place. Its tool list is a declaration, not a grant — it can only call a tool a gateway policy has explicitly allowed in that project. Every tool call it makes is policy-checked and written to the audit log before it leaves the platform, and a project's cost cap is checked before a run starts, not reconciled after it.
More in Development
- D1Frontend Web Development AgentImplements web UI in code (Next.js/React) from design specs and design-system tokens.Read the product page
- D2Mobile Development AgentImplements iOS/Android (native or React Native/Flutter) screens and platform-specific behavior.Read the product page
- D3Backend Development AgentImplements business logic, services, and API endpoints (FastAPI/async) per architecture contracts.Read the product page
- D4Database Development AgentDesigns and evolves schema, migrations, indexing, and query optimization (PostgreSQL/pgvector).Read the product page
- D5Third-Party Integration AgentConnects agents and services to external systems (payments, email, analytics, auth providers) — increasingly via existing MCP servers rather than hand-built connectors.Read the product page
- D6Localization/Internationalization AgentExtracts translatable UI strings, manages locale files, and adapts formatting (dates, currency, RTL layout) for each target market.Read the product page
- D7Data Migration AgentPlans and executes one-time bulk data migrations when onboarding existing/legacy applications — schema mapping, data validation, cutover strategy.Read the product page
- D8Data Engineering / Analytics Engineering AgentBuilds and maintains ETL/ELT pipelines, the analytics data warehouse, and data-quality checks that move data from operational services into a form Product Analytics Agent and FinOps/Cost Optimization Agent can query; self-heals pipelines when a source schema drifts.Read the product page