The all-in-one AI plugin for AI Engineers

Skills for ai engineers 17

• code debugging assistantTraces a failing pytest run to the bad import and returns a fixed file plus a regression test.
• agency engineering ai engineerTakes a churn model from notebook to a served endpoint with data pipeline and integration steps listed.
• senior ml engineerPlans a RAG service with chunking, vector store, evaluation, drift alerts and rollback for your docs.
• documentationTurns your inference API into a README and endpoint reference with request examples and error codes.
• feature engineering advisorFixes a skewed price column with log transform, imputation and encoding before you retrain the model.
• ml algorithm selection assistantCompares gradient boosting and logistic regression on your 40k row churn set, with metrics and tuning plan.

Plus a start-here assistant, the prompt library and the app builder, and for members the daily brief.

Ready-made prompts 25

• Code Drafting Basics3 prompts, like Generate ML Boilerplate Code
• Debugging and Errors3 prompts, like Error Message Explanation and Resolution
• Explaining Concepts and Papers3 prompts, like Summarize a Research Paper
• Data Preparation and Exploration4 prompts, like Plan Dataset Cleaning Steps
• Model Training and Tuning3 prompts, like Design Training Loop Components
• Deployment and Production3 prompts, like Draft Docker and API Serving Code
• Monitoring and Optimization3 prompts, like Set Up Production Model Monitoring Metrics

Apps you can build with AI 8

• Managed model training and deployment workspaceKeeps training runs, artifacts and deploys in one place so your team stops hand-copying models between tools. Demo included.
• AI model test and monitoring workbenchGives you one defensible record of model tests and live metrics before each release. Demo included.
• Prompt lifecycle testing and deployment workbenchTracks every prompt version and test result so a prompt change does not quietly break production. Demo included.
• Raw-source data preparation and stewardship libraryCuts repeated data cleaning by keeping source rights, review notes and prepared sets in one library. Demo included.
• Production agent trace and evaluation consoleShows where your agent failed in production and scores quality drift without shipping traces to a third party. Demo included.
• Source-linked multi-model text and code workbenchRuns generation, coding and analysis across models in one console with outputs linked to their sources. Demo included.
• Source-linked code review and debugging consoleShortens review cycles while keeping a named engineer's approval on every merged change. Demo included.
• Source-linked language model workbenchKeeps prompts, context and data in one owned place instead of scattered across vendor tools. Demo included.

AI agents 12

In the download. An agent runs a whole recurring task for you, step by step, and asks you before anything important. Ask your AI: "Run the Build Environment Repair Agent". Each one also has a guided set-up with sample data on its page.

• Build Environment Repair AgentA reproducible build environment documented for everyone.
• Product Documentation Task Test AgentDocs examples that work exactly as written.
• Experiment Reproducibility Packaging AgentA tested reproduction bundle, with any remaining failures stated.
• Model Fairness Audit AgentA documented review of how a model performs across groups, with tested mitigations and tradeoffs
• Label Quality Audit AgentFind and fix the label errors that hold model quality down, and prove the fix with a retrained comparison
• Data Leakage Detection AgentRemove leaks from splits and features and produce an honest performance estimate
• Champion-Challenger Rollout AgentMove a replacement model from shadow to full traffic only while the evidence supports each step
• Decision Threshold Tuning AgentSet decision thresholds that minimize real-world cost and stay stable as data shifts
• Inference Cost and Latency Review AgentCut serving cost and tail latency without lowering model quality
• Prompt Injection Red-Team AgentFind the ways your prompt can be attacked and fix them without hurting normal use
• Golden Test Set Maintenance AgentKeep the test set current so real failures become permanent tests
• Model Selection Bake-Off AgentChoose a model with evidence on quality, cost and speed for the real task

Connections (MCP servers) 12

A guide in the download. A connection lets your AI work in an app you already use. You add only the ones you use: ask your AI "Which apps can you connect to?" and follow the steps on each connection's page.

• Jungle Grid MCP serverLets your AI estimate, run and check GPU jobs on Jungle Grid, then fetch the results. Official. No install needed.
• TurboQuant MCP serverLet your AI shrink big language models into smaller files and send them to Hugging Face.
• MLflow MCP serverLets your AI look up MLflow experiments, runs, models, and traces and summarize them for you.
• Google AI Search MCP serverLets your AI search the web and pull answers from official docs using Google's AI models.
• Tech Trend MCP serverYour AI can gather fresh research papers, trending code, and FDA device news, then write you a short newspaper-style briefing.
• Jupyter MCP serverLet your AI open, edit and run Jupyter notebooks for you while you watch. No install needed.
• Optuna MCP serverLet your AI run optimization experiments and show you the results as charts. Official.
• Papers with Code MCP serverLets your AI search research papers and find the code, datasets, and results linked to them.
• Google Colab MCP serverLet your AI open a Google Colab notebook, add and run code cells, and pick a GPU for you.
• pcq MCP serverLets your AI run, check and compare machine learning experiments in your project and report back what happened.
• Jupyter Notebook MCP serverLet your AI write and run code inside your Jupyter Notebook, cell by cell, and read the results.
• arXiv MCP serverLets your AI search arXiv for research papers and pull up their details and full text. No install needed.

Your Grok Bot team 12

• Ml EngineerBuild and maintain production ML systems with PyTorch, TensorFlow, and modern MLOps practices.
• Pytorch LightningOrganize PyTorch code into LightningModules and configure Trainers for scalable neural network training.
• Python DevelopmentScaffold production-ready Python projects with modern tooling and type hints.
• Deployment EngineerDesigns and optimizes CI/CD pipelines for faster, safer deployments with automated rollbacks and monitoring, including GitOps and progressive delivery
• Distributed Training Ray TrainScales PyTorch, TensorFlow, and HuggingFace training from one GPU to thousands of nodes across a cluster.
• Data Processing Ray DataProcess large ML datasets with Ray Data across CPU/GPU clusters.
• Remote Gpu TrainerDeploy, monitor, and debug long GPU jobs on rented instances with safe teardown and resumable checkpoints.
• Ml Pipeline WorkflowEnd-to-end MLOps pipeline orchestration from data prep to model deployment and monitoring.
• Machine Learning EngineerDeploy and optimize ML models for production inference at scale.
• Python Performance OptimizationProfile and optimize Python code for speed and memory efficiency.
• Test Driven DevelopmentEnforce the TDD cycle: write failing test first, minimal code to pass, verify both steps.
• Model Architecture TorchtitanPretrains large language models at scale using PyTorch-native torchtitan with 4D parallelism.