Loading...
Loading...
Found 31 Skills
Train ML models on Databricks. Use for: classification/regression/deep-learning (XGBoost, scikit-learn, LightGBM, PyTorch) with Optuna, @prod/@challenger aliases, batch scoring (spark_udf for plain models, fe.score_batch for feature-store-backed), custom PyFunc, custom ResponsesAgent (LangGraph + UC Function/Vector Search); UC feature tables + FeatureLookup + point-in-time joins + Lakebase online store; declarative Feature Views (create_feature, DeltaTableSource, RollingWindow/SlidingWindow/TumblingWindow, materialize_features, streaming Kafka features). NOT for: endpoint ops (databricks-model-serving), MLflow evaluation (databricks-mlflow-evaluation).
Use to help users get started with Nemo Gym reward profiling. Covers the basic ng_run, ng_collect_rollouts, and ng_reward_profile workflow, repeated rollouts, materialized inputs, rollout JSONL artifacts, task and rollout identity, output inspection, partial profiling, and rollout_infos. For failed jobs, prefer nemo-gym-debugging.
Design the structural anatomy of screens at wireframe fidelity — what goes where and why, before anyone argues about how it looks. Part of the Intent design strategy system. Produces lo-fi idea boards for divergent exploration, complete interactive grayscale wireframes with real labels and real hierarchy, and click-through prototypes that materialize flow logic from /journey. Trigger on: wireframe, wireframes, wireframing, thumbnails, "sketch the screen", "lay out this page", "what goes where on this screen", screen layout, page structure, lo-fi, mid-fi, click-through prototype, wireflow, "wireframe the dashboard", or any request to design the structure of a screen before its visual design. The flow through screens belongs to /journey; the information structure belongs to /organize; the words belong to /articulate — this skill owns the screen itself.
Apache Iceberg tables on Databricks — Managed Iceberg tables, External Iceberg Reads (fka Uniform), Compatibility Mode, Iceberg REST Catalog (IRC), Iceberg v3, Snowflake interop, PyIceberg, OSS Spark, external engine access and credential vending. Use when creating Iceberg tables, enabling External Iceberg Reads (uniform) on Delta tables (including Streaming Tables and Materialized Views via compatibility mode), configuring external engines to read Databricks tables via Unity Catalog IRC, integrating with Snowflake catalog to read Foreign Iceberg tables
Guidance for detection engineering in Microsoft Sentinel — building, testing, deploying, and maintaining analytics rules, hunting queries, and SOAR automation. Covers the Content Hub solution model, MITRE ATT&CK mapping, scheduled vs near-real-time (NRT) vs Fusion vs anomalies analytics, KQL detection patterns (joins, summarize, bin, materialize), entity mapping and incident enrichment, custom detections from Defender XDR vs Sentinel-only, automation rules, playbooks (Logic Apps), watchlists, threat intel matching, content as code with Azure DevOps / GitHub repositories integration, and detection lifecycle (validate → tune → version). WHEN: Sentinel analytics rule, KQL detection, MITRE mapping, Sentinel content hub, scheduled analytics, NRT rule, hunting query, Sentinel automation rule, Logic App playbook, custom detection, repositories Sentinel CI/CD, detection-as-code, watchlist, threat intel matching analytics, fusion alerts, anomalies, incident enrichment, entity mapping. DO NOT USE for Sentinel architecture/onboarding (use sentinel), Defender XDR custom detections only (overlap—use the side that owns the data), or generic KQL training.
Diagnose ClickHouse Kafka engine health, consumer status, thread pool capacity, and consumption issues. Use for Kafka lag, consumer errors, and thread starvation.
Diagnose ClickHouse INSERT performance, batch sizing, part creation patterns, and ingestion bottlenecks. Use for slow inserts and data pipeline issues.