Skill UI

Guides building production-grade LLM-as-judge evaluation systems by defining scoring and pairwise comparison setups, bias mitigation, and metric choices for automated and human-aligned quality assessment.

Simulates Andrej Karpathy to coach deep learning, explain LLMs, Software 2.0, and autonomous systems with practical neural net builds, ideal when users seek Karpathy-style pedagogy or deep technical clarity.

Assess how likely AI systems such as ChatGPT, Claude, Perplexity or Gemini are to extract and cite content from a page by scoring answer blocks, self-contained facts, and structure, and by giving rewrite tips to improve visibility in AI responses.

Generates and validates llms.txt files, the emerging standard that tells AI systems how to prioritize and understand a website; crawls content, ranks pages, and enforces formatting so your brand narrative stays consistent.

Deploys machine-learning anomaly detection across OT networks by modeling deterministic SCADA/Modbus/DNP3/OPC UA traffic, correlating with historians, and flagging behavior deviations, rogue devices, or process anomalies for industrial control security monitoring.

Builds NeMo Guardrails and custom Python validators to enforce input/output safety for LLM apps, blocking prompt injection, PII leaks, toxic outputs, hallucinations, and schema violations before they reach downstream systems.