Luniat Magazine https://luniat.com/en/magazine A handbook in five parts on going from scattered AI experiments to systems in production, with practical guides per industry for smaller businesses. en-GB Fri, 02 Oct 2026 08:00:00 GMT Agents in production: state machines, budgets and humans in the loop https://luniat.com/en/magazine/engineering/agents-in-production https://luniat.com/en/magazine/engineering/agents-in-production Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 10 Run an agent as an explicit state machine with a step budget, a deadline and an error cap, saved after every step and paused for human approval. Cost and latency: caching, batching, routing and budgets https://luniat.com/en/magazine/engineering/cost-and-latency https://luniat.com/en/magazine/engineering/cost-and-latency Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 07 Where the money and the milliseconds go in an LLM system, and four levers that move both, with the arithmetic and tested code for each. Evals as CI: release gates for LLM systems https://luniat.com/en/magazine/engineering/evals-as-ci https://luniat.com/en/magazine/engineering/evals-as-ci Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 02 Versioned datasets, deterministic scorers, a calibrated LLM judge and a paired statistical gate that blocks a merge when quality drops. With tested code. One rate limit, many tenants: fair queues and adaptive throttling https://luniat.com/en/magazine/engineering/fair-scheduling-and-rate-limits https://luniat.com/en/magazine/engineering/fair-scheduling-and-rate-limits Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 06 Share one provider rate limit between tenants with deficit round robin, pace calls from quota headers, and keep a reserve for interactive traffic. A reference architecture for LLM systems in production https://luniat.com/en/magazine/engineering/llm-gateway-reference-architecture https://luniat.com/en/magazine/engineering/llm-gateway-reference-architecture Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 01 One gateway between your services and every model provider, with per-tenant limits, versioned prompts, fallbacks, retries and telemetry. Tested code. Security for LLM applications: injection, exfiltration and excessive agency https://luniat.com/en/magazine/engineering/llm-security-in-production https://luniat.com/en/magazine/engineering/llm-security-in-production Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 08 Why prompt injection cannot be filtered away, and the controls that hold anyway: spotlighting, taint-aware tool policies, output sanitising. Multi-tenant AI without leaks: isolation from the database to the prompt https://luniat.com/en/magazine/engineering/multi-tenant-ai https://luniat.com/en/magazine/engineering/multi-tenant-ai Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 05 Forced row-level security tested in CI, a request-scoped tenant context, tenant-bound caches and secrets, and why the prompt is a new place to leak data. Observability for LLM systems: traces, metrics and samples without leaking data https://luniat.com/en/magazine/engineering/observability-for-llm-systems https://luniat.com/en/magazine/engineering/observability-for-llm-systems Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 09 Spans with OpenTelemetry's generative AI attributes, tail sampling that keeps every interesting trace, latency histograms, and no prompts in the logs. Retrieval that holds up: hybrid search, chunking and citations in Postgres https://luniat.com/en/magazine/engineering/rag-in-production https://luniat.com/en/magazine/engineering/rag-in-production Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 03 Structure-aware chunking, vector and full-text search fused with RRF in one SQL query, tenant-scoped by construction, with citations you can verify. Structured output and tool calls that do not break production https://luniat.com/en/magazine/engineering/structured-output-and-tools https://luniat.com/en/magazine/engineering/structured-output-and-tools Fri, 02 Oct 2026 08:00:00 GMT Engineering · No. 04 Validate model output against a schema, repair it with exact errors, and run tool calls with permissions, validation and idempotency keys. Tested code.