Notes on ML systems, AI platforms and agents.https://www.ammaradam.dev/https://www.ammaradam.dev/posts/patterns-for-safe-tool-calls/https://www.ammaradam.dev/posts/patterns-for-safe-tool-calls/Most of what makes an agent safe to run happens at the tool call, not in the prompt: who the call acts for, how risky it is, what happens when it is retried, and whether anyone can see it afterwards.Mon, 19 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/hide-complexity-not-evidence/https://www.ammaradam.dev/posts/hide-complexity-not-evidence/An abstraction can reduce how much infrastructure a user needs to know without reducing the evidence they need to diagnose a problem.Thu, 15 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/agent-identity-is-not-authorization/https://www.ammaradam.dev/posts/agent-identity-is-not-authorization/Knowing who an agent is answers only part of the question. Whether an action is allowed depends on the user, the scope they delegated and the resource's own policy.Tue, 13 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/time-to-first-production-call/https://www.ammaradam.dev/posts/time-to-first-production-call/Platform metrics are more useful when they cover how quickly a new team gets started, not only how the platform performs once it's running.Thu, 08 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/gateway-errors-that-help/https://www.ammaradam.dev/posts/gateway-errors-that-help/A blocked request is more useful when the error says what happened, where, and what to try next, without exposing sensitive details.Tue, 06 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/standardise-the-right-things/https://www.ammaradam.dev/posts/standardise-the-right-things/Shared platforms tend to work best when they standardise the capabilities every team would otherwise rebuild, and leave use-case decisions to the teams.Thu, 01 Oct 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/first-users-shape-the-design/https://www.ammaradam.dev/posts/first-users-shape-the-design/A platform's first real users show where its abstractions help and where they get in the way, which makes their feedback part of the design work.Tue, 29 Sep 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/who-chooses-the-model/https://www.ammaradam.dev/posts/who-chooses-the-model/Pinning a model version gives reproducibility; letting a platform route by requirements eases upgrades. Model retirement schedules make the trade-off concrete.Thu, 24 Sep 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/tool-call-timeouts/https://www.ammaradam.dev/posts/tool-call-timeouts/A timed-out tool call may still have completed, so retrying safely needs the familiar distributed-systems tools: idempotency keys, durable state and verification.Tue, 22 Sep 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/agent-success-is-not-a-200/https://www.ammaradam.dev/posts/agent-success-is-not-a-200/Infrastructure can be healthy while an agent's task goes wrong, so agent observability leans on execution traces, not only logs and latency.Fri, 18 Sep 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/checking-user-and-agent/https://www.ammaradam.dev/posts/checking-user-and-agent/Workload identity says which software is calling; authorizing an agent's action also needs to know which user it acts for and through which chain of agents.Wed, 16 Sep 2026 00:00:00 GMThttps://www.ammaradam.dev/posts/what-not-to-expose/https://www.ammaradam.dev/posts/what-not-to-expose/Infrastructure was largely designed for people who read docs and debug by hand; building for agents and for non-infrastructure users both raise the question of what to hide.Mon, 14 Sep 2026 00:00:00 GMT