Yishto — AI engineering guideshttps://yishto.com/en/learn/From a model call to a dependable AI app.enUnderstand context before your first model callhttps://yishto.com/en/learn/models/context-and-tokens/https://yishto.com/en/learn/models/context-and-tokens/Separate tokens, context windows, and output budgets, then practice assembling model input with an offline example.Mon, 09 Nov 2026 16:00:00 GMTStreaming and structured output solve different problemshttps://yishto.com/en/learn/models/streaming-and-structured-output/https://yishto.com/en/learn/models/streaming-and-structured-output/Separate incremental display from a complete business result and validate output with a local Python example.Mon, 09 Nov 2026 16:00:00 GMTBuild a task evaluation set before choosing a modelhttps://yishto.com/en/learn/models/model-evaluation/https://yishto.com/en/learn/models/model-evaluation/Define success with actual tasks and compare quality, latency, and failure modes rather than a leaderboard alone.Mon, 09 Nov 2026 16:00:00 GMTThe smallest RAG loop: retrieve, cite, answerhttps://yishto.com/en/learn/rag/retrieval-pipeline/https://yishto.com/en/learn/rag/retrieval-pipeline/Separate retrieval from generation and use offline keyword retrieval to understand provenance and missing evidence.Mon, 09 Nov 2026 16:00:00 GMTChunk documents for meaning before tuning an indexhttps://yishto.com/en/learn/rag/chunking-and-indexes/https://yishto.com/en/learn/rag/chunking-and-indexes/Understand chunking, overlap, metadata, and approximate vector indexes using a small windowing example.Mon, 09 Nov 2026 16:00:00 GMTEvaluate retrieval and answers separately in RAGhttps://yishto.com/en/learn/rag/rag-evaluation/https://yishto.com/en/learn/rag/rag-evaluation/Use Recall@k, evidence support, and unanswerable cases to locate failures in a RAG pipeline.Mon, 09 Nov 2026 16:00:00 GMTWhen an agent helps and when a workflow is enoughhttps://yishto.com/en/learn/agents/workflows-and-agents/https://yishto.com/en/learn/agents/workflows-and-agents/Decide using control flow, task uncertainty, and failure cost, then separate model decisions from program rules.Mon, 09 Nov 2026 16:00:00 GMTTool calling and MCP still need execution boundarieshttps://yishto.com/en/learn/agents/tool-contracts-and-mcp/https://yishto.com/en/learn/agents/tool-contracts-and-mcp/Define input, output, error, and permission contracts and practice with an offline tool dispatcher.Mon, 09 Nov 2026 16:00:00 GMTRecoverable agents: state, retries, and idempotencyhttps://yishto.com/en/learn/agents/durable-agent-state/https://yishto.com/en/learn/agents/durable-agent-state/Use SQLite to practice durable task transitions and distinguish conversation history from external side effects.Mon, 09 Nov 2026 16:00:00 GMTPrompt injection: keep untrusted content inside data boundarieshttps://yishto.com/en/learn/production/prompt-injection/https://yishto.com/en/learn/production/prompt-injection/Recognize indirect instructions in documents and tool results and enforce permissions in application code.Mon, 09 Nov 2026 16:00:00 GMTObserve AI apps: first-token latency, total time, and costhttps://yishto.com/en/learn/production/latency-cost-observability/https://yishto.com/en/learn/production/latency-cost-observability/Separate retrieval, model, and tool latency, calculate percentiles, and record request-level telemetry.Mon, 09 Nov 2026 16:00:00 GMTFrom demo to deployment: design the failure pathshttps://yishto.com/en/learn/production/shipping-checklist/https://yishto.com/en/learn/production/shipping-checklist/Check timeouts, rate limits, missing evidence, and tool failures using a practical deployment matrix.Mon, 09 Nov 2026 16:00:00 GMT