Skill UI

nanoGPT is a simplified, minimalist implementation designed specifically for educational purposes. It reproduces the core architecture of GPT-2 (124M) using clean, hackable code, allowing users to understand the entire transformer pipeline from scratch. The framework supports full workflows, including data preparation on various datasets (e.g., Shakespeare, OpenWebText), model training, and text generation, making it ideal for students and researchers learning NLP and deep learning principles.

NeMo Curator streams GPU-accelerated curation for LLM training, with super-fast fuzzy deduplication, multimodal quality filters, semantic dedupl, PII redaction, NSFW checks, and RAPIDS-scaled pipelines for prepping massive corpora.

Automates Neon serverless Postgres workflows via the Composio MCP, letting you list projects and branches, inspect databases, manage roles, and retrieve connection URIs for pipelines and tooling integrations.

Monitors Nx Cloud CI pipelines by running the ci-watcher subagent, polling execution status, and applying self-healing fixes or rerun requests while honoring user overrides and Nx Cloud connection requirements.

Automates onnx converter work for ML deployment, providing best-practice guidance on serving, pipelines, monitoring, and optimization while generating production-ready code and configs.

Integrates OpenEvidence clinical AI testing into CI/CD pipelines by orchestrating GitHub Actions workflows, vitest unit/integration suites, and clinical validation checks so teams can automate clinical query verification with sandbox credentials and structured retries.

Orchestrates complex multi-stage deployment pipelines by generating production-ready configs and setup code based on user requirements, ideal for automating CI/CD and reliable software releases.

Coordinates parallel test execution across suites, frameworks, and CI workers, handling sharding, retries, report aggregation, and resource-aware tiers for faster, reliable pipelines.

Pandas Pro orchestrates DataFrame transformations for cleaning, aggregating, merging, resampling, and optimizing large tables using vectorized patterns and validation checks, ideal for analytics pipelines.

Comprehensive PDF toolkit for extracting text and tables, splitting/merging documents, filling forms, and generating new files via Python or CLI tools, enabling automated, large-scale document workflows in analytics, compliance, or reporting pipelines.

Guides choosing and implementing Perplexity-validated architectures across scales, from simple widgets to cached layers and automated research pipelines, including deployment trade-offs and error mitigation.

Defines a Perplexity reference architecture for AI research and search projects, covering model routing, citation extraction, conversational search, and multi-query pipelines to establish consistent integrations and reviews.