<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>Joshua Fields - Writing</title>
    <link>https://joshuafields.uk/blog</link>
    <atom:link href="https://joshuafields.uk/feed.xml" rel="self" type="application/rss+xml"/>
    <description>Technical writing by Joshua Fields, a London-based AI and software engineer, on AI agents, authorization, real-time voice AI and full-stack engineering.</description>
    <language>en-gb</language>
    <lastBuildDate>Sun, 25 Oct 2026 09:00:00 GMT</lastBuildDate>
    <item>
      <title>Authorizing AI Agents: Zanzibar-Style Permissions for MCP Tool Calls</title>
      <link>https://joshuafields.uk/blog/authorizing-ai-agents-mcp-tool-calls-zanzibar</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/authorizing-ai-agents-mcp-tool-calls-zanzibar</guid>
      <pubDate>Sun, 25 Oct 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields, a London-based AI engineer, shows how to authorize AI agent and MCP tool calls with Zanzibar-style ReBAC: delegation, tool grants and per-resource checks in one schema, with transactional revocation and measured latency using pemba.</description>
      <category>AI engineer</category>
      <category>AI agents</category>
      <category>MCP</category>
      <category>Model Context Protocol</category>
      <category>agent authorization</category>
      <category>Zanzibar authorization</category>
      <category>ReBAC</category>
      <category>SpiceDB</category>
      <category>Python</category>
      <category>AI security</category>
      <category>London</category>
    </item>
    <item>
      <title>Building pemba: An Embedded Zanzibar-Style Authorization Engine in Rust</title>
      <link>https://joshuafields.uk/blog/building-pemba-embedded-zanzibar-authorization-rust</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/building-pemba-embedded-zanzibar-authorization-rust</guid>
      <pubDate>Mon, 19 Oct 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields, a London-based software engineer, on building pemba: an embedded, Zanzibar-style authorization engine in Rust and Python whose permission writes join the application&apos;s own SQLite transaction, with SCC fixed-point evaluation for cyclic data and differential testing against SpiceDB.</description>
      <category>software engineer</category>
      <category>Rust</category>
      <category>Python</category>
      <category>Zanzibar authorization</category>
      <category>SpiceDB</category>
      <category>SQLite</category>
      <category>property-based testing</category>
      <category>open source</category>
      <category>London</category>
    </item>
    <item>
      <title>Building Authorization Infrastructure with SpiceDB, Go and Kubernetes</title>
      <link>https://joshuafields.uk/blog/building-authorization-infrastructure-spicedb-go-kubernetes</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/building-authorization-infrastructure-spicedb-go-kubernetes</guid>
      <pubDate>Mon, 22 Jun 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields, a London-based software engineer, explains how to think about authorization infrastructure using SpiceDB, Zanzibar-style permissions, Go services, Kubernetes, schema-as-code, and production rollout strategy.</description>
      <category>software engineer</category>
      <category>SpiceDB</category>
      <category>Zanzibar authorization</category>
      <category>Go</category>
      <category>Kubernetes</category>
      <category>backend infrastructure</category>
      <category>AI infrastructure</category>
      <category>London</category>
    </item>
    <item>
      <title>Building Real-Time Voice AI Applications with LiveKit and FastAPI</title>
      <link>https://joshuafields.uk/blog/building-real-time-voice-ai-applications-livekit-fastapi</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/building-real-time-voice-ai-applications-livekit-fastapi</guid>
      <pubDate>Mon, 22 Jun 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields, a London-based software engineer and creative technologist, explains how real-time voice AI applications can be designed with LiveKit, FastAPI, LLMs, speech-to-text, text-to-speech, and production-grade backend architecture.</description>
      <category>software engineer</category>
      <category>London</category>
      <category>real-time voice AI</category>
      <category>LiveKit</category>
      <category>FastAPI</category>
      <category>LLM applications</category>
      <category>AI infrastructure</category>
    </item>
    <item>
      <title>Designing Explainable AI Matchmaking Systems</title>
      <link>https://joshuafields.uk/blog/designing-explainable-ai-matchmaking-systems</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/designing-explainable-ai-matchmaking-systems</guid>
      <pubDate>Mon, 22 Jun 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields writes about building explainable AI-assisted matchmaking systems using conversational onboarding, profile enrichment, deterministic scoring, privacy-aware design, and human-centred product thinking.</description>
      <category>AI matchmaking</category>
      <category>software engineer</category>
      <category>creative technologist</category>
      <category>explainable AI</category>
      <category>London</category>
      <category>full-stack AI product</category>
    </item>
    <item>
      <title>When Should Governments Pull the Plug on AI Models?</title>
      <link>https://joshuafields.uk/blog/when-should-governments-pull-the-plug-on-ai-models</link>
      <guid isPermaLink="true">https://joshuafields.uk/blog/when-should-governments-pull-the-plug-on-ai-models</guid>
      <pubDate>Mon, 22 Jun 2026 09:00:00 GMT</pubDate>
      <dc:creator>Joshua Fields</dc:creator>
      <description>Joshua Fields examines when government intervention in frontier AI may be justified, and why capability-based controls, access design, and infrastructure-level accountability matter for safe deployment.</description>
      <category>AI policy</category>
      <category>software engineering</category>
      <category>AI infrastructure</category>
      <category>cybersecurity</category>
      <category>access control</category>
    </item>
  </channel>
</rss>
