Building Agents

Answer questions from a curated index with citations to source documents

Answer real-time questions with tool calls for time, weather and unit conversion

Customize a Nemotron voice agent for healthcare appointments and intake

Build and review NeMo Fabric adapters for third-party agent harnesses and runtimes

Integrate the NeMo Fabric Python SDK into an app, service or eval harness

Debug missing events and load failures in NeMo Relay runtime integrations

Choose and follow a NeMo Relay quick start and verify the first integration

Install NeMo Relay for an agent runtime, language or plugin bundle

Wrap LLM and tool calls with NeMo Relay scopes for events and guardrails

Keep NeMo Relay scope stacks separate across async tasks, threads and workers

Add typed wrappers and provider codecs to NeMo Relay without changing behavior

Tune NeMo Relay plugin telemetry, adaptive hints and tool parallelism

Package reusable NeMo Relay behavior as embedded or native plugins

Set up NeMo Relay observability with OpenTelemetry, OpenInference or ATIF

Search and ingest PDFs, Office files and media with the NeMo Retriever CLI

Generate custom safety policies and taxonomies for Nemotron guardrails

Plan, tune and deploy Nemotron embed and rerank retrieval recipes

Build and fix real-time voice agents with Pipecat or LiveKit pipelines

Deploy, configure and troubleshoot the NVIDIA RAG Blueprint and its services

Score RAG answer quality with RAGAS on a corpus and training questions

Load-test a deployed RAG Blueprint server with profiling and aiperf from a YAML config

Add captioning, detection, search and alerts to an app using the VSS Blueprint

Call OpenRouter LLM APIs with streaming and web search from a PinMe Worker

Audit local coding agent sessions for cost, tokens, tool failures and latency

Run CIAgent regression checks after agent code or prompt changes and read results

Write an AI design contract and framework guidance for a phase that builds AI features

Audit eval coverage for an executed AI phase and list remediation steps

Set up CIAgent regression tests with golden baselines for an AI agent repo

Score RAG retrieval and citation quality from JSONL fixtures in offline evals

Route AI model calls through RouterBase with fallbacks and chosen model IDs

Operate and debug a GoClaw gateway with its CLI for agents, tools and channels

Design metrics, calibration and eval suites that show whether an agent really improved

Choose an AWS runtime for an AI agent and plan a proof of concept

Extend an AgentCore agent project with memory, VPC, multi-agent and browser tools

Diagnose broken AgentCore agents from traces and logs, including tool call failures

Scaffold a new AgentCore agent project from framework choice to first invocation

Harden an AgentCore agent for production with IAM scoping, auth and rate limits

Add evaluators, online monitoring and quality gates to measure and speed up an agent

Let an agent pay x402 paywalls at runtime within operator spend limits

Build generative AI apps on Amazon Bedrock with Knowledge Bases and Agents

Migrate OpenAI, Gemini and Anthropic API code to Amazon Bedrock and test output

Diagnose one agent run step by step in an incident-style canvas report

Let LLMs run generated TypeScript in sandboxed isolates with snippet libraries

Route to TanStack AI guides for chat, tools, media, outputs and adapters

Configure TanStack AI provider adapters like OpenAI, Anthropic, Gemini and Ollama

Stream AG-UI protocol events from a TanStack AI server over SSE or NDJSON

Build TanStack AI chat endpoints and streaming UIs with useChat and message parts

Generate images, speech, video and transcripts with TanStack AI media adapters

Add chat lifecycle hooks for analytics, logging, tool caching and tracing