AI, RAG & Vector Search
This hub explains how to build AI systems that can retrieve evidence, isolate tenant data, control model behaviour and measure answer quality—not merely call an LLM API.
When should Qdrant, PostgreSQL search or SQL be used?
How are groundedness, citations and tool use evaluated?
Define the boundary, then choose the technology.
Use RAG
Knowledge changes, evidence matters and the answer must cite private or domain-specific sources.
Use search or SQL
The user needs exact records, aggregates, filters or deterministic business facts.
Add an agent
The system must choose tools or execute a bounded workflow with authorization and audit evidence.
Continue with practical guides.
GPT-Image 2.6: build a production asset pipeline, not a prompt demo
OpenAI's GPT-Image 2.6 adds Sunburst, Flare, complexity, markup edits, SVG and GIF. Build a governed asset pipeline with validation, evaluation and rollback.
Read the guideAI · Agents · Model routingClaude Fable 5.2: route frontier work without making it the default
Migrate to Claude Fable 5.2 with workload routing, retention gates, tool compatibility, evals, canaries and a clean rollback plan.
Read the guideSaudi Technology · AI Governance · Media EngineeringSaudi AI media principles: build a disclosure gate
Turn Saudi Arabia's eight AI media principles into a production gate for disclosure, verification, privacy, provenance and auditable publishing.
Read the guideAI engineering · Cybersecurity · AgentsGPT-6 Cyber: build an authorized research control plane
A production guide to GPT-6-Cyber: Daybreak Red access, scope manifests, isolated labs, tool approvals, evidence capture and independent patch verification.
Read the guideProduct engineering · AI agentsAgent tool context is a product budget, not a prompt dump
A Product Engineering framework for controlling AI-agent tool definitions, deferred discovery, result compaction and caching against user outcomes, latency, safety and cost per successful task.
Read the guideOpenAI · Coding agents · Distributed systemsOpenAI Pando: design Git infrastructure for coding-agent scale
A source-backed analysis of OpenAI's Pando architecture and a practical design for source control that stays correct under coding-agent concurrency.
Read the guideSaudi technology · AI infrastructure · CloudAramco takes 49% of HUMAIN: what Saudi AI teams should verify
A source-backed analysis of the closed PIF–Aramco HUMAIN transaction and an engineering checklist for evaluating Saudi full-stack AI infrastructure.
Read the guideAI engineering · Anthropic · API migrationClaude Opus 4.8 fast mode retires: migrate without latency regressions
A production migration guide for removing Claude Opus 4.8 fast-mode dependencies without confusing feature retirement with model retirement or hiding compatibility regressions.
Read the guideOpenAI · Life sciences AI · GovernanceGPT-6 Rosalind Research: build a governed life-sciences workflow
A production guide to GPT-6 Rosalind Research access, data boundaries, scientific tools, evidence, human review, migration, audit and revocation.
Read the guideOpenAI · AI agents · GovernanceOpenAI Model Spec: turn agent autonomy into an action contract
A production guide to translating OpenAI's updated guidance on intent, context, agency and side effects into deterministic controls around tool-using agents.
Read the guideAnthropic · Claude Enterprise · Slack agentsClaude Tag API: govern Slack agents as code
A source-backed guide to governing Claude in Slack with desired state, least privilege, bundle controls and verified reconciliation.
Read the guideSaudi technology · AI safety · DeepfakesSaudi deepfake campaign: build verification into products
A source-backed analysis of Saudi Arabia’s deepfake campaign and an auditable media-verification design for product and commerce teams.
Read the guide