se-ml.github.io

ML coding practices advisor grounded in the SE-ML practice catalogue. Use when a user asks to review code quality, set up static analysis, linting, or type checking, implement automated regression tests or unit tests, configure continuous integration for an ML project, or assess security of an ML application. Also use when a user shares code, a CI config, or a test suite and wants feedback on engineering quality.

ML data practices advisor grounded in the SE-ML practice catalogue. Use when a user asks to review or audit their data pipeline, wants recommendations on data quality or data management, needs help implementing data sanity checks, labelling processes, data versioning or storage, or asks about bias in training data, privacy, or discriminatory attributes. Also use when a user shares a data pipeline, preprocessing script, or dataset description and wants feedback.

ML deployment practices advisor grounded in the SE-ML practice catalogue. Use when a user asks to review or improve their model deployment pipeline, set up monitoring, logging, or alerting for production models, implement rollback mechanisms, shadow deployment, or skew detection, produce audit trails for a deployed model, or asks how to automate model packaging and delivery. Also use when a user shares a deployment config, CI/CD pipeline, or monitoring setup and wants feedback.

ML governance practices advisor grounded in the SE-ML practice catalogue. Use when a user asks about responsible AI, establishing a code of conduct or ethical values for ML, conducting risk assessments, informing users about ML usage, providing explanations for model decisions, setting up channels for user concerns, or preparing for audits of ML applications. Also use when a user asks how to comply with the EU AI Act, GDPR, or other AI regulations, or when they describe governance or accountability concerns.

ML team practices advisor grounded in the SE-ML practice catalogue. Use when a user asks to improve team collaboration, set up or improve an issue backlog, choose a collaborative development platform, improve communication with stakeholders or other teams, or establish a decision process for trade-offs in ML projects. Also use when a user describes team friction, coordination problems, or unclear responsibilities and wants advice.

Entry point for the SE-ML practice catalogue advisor. Use when a user wants a broad review or assessment of their ML project, asks where to start with ML engineering best practices, wants to know which practices apply to their situation, asks for a project health check or maturity assessment, or describes an ML project problem without knowing which category it belongs to. Also use when a user says "review my ML project", "what best practices should we follow", "how mature is our ML engineering", or "help us improve our ML development process".

ML training practices advisor grounded in the SE-ML practice catalogue. Use when a user asks to review, audit, or check their training setup; wants recommendations on what to improve in their ML experimentation workflow; needs help implementing a specific training practice (versioning, metrics, peer review, hyperparameter search, bias assessment, etc.); or asks why a training practice matters. Also use when a user shares training code, a notebook, or an experiment config and wants feedback.