A 12-week hands-on AI engineering mentorship. Build a real AI coding agent from scratch — agent loops, tool calling, evals, RAG, context engineering, sandboxing, memory and multi-agent orchestration. No frameworks.https://pjsofts.github.io/en-ushttps://pjsofts.github.io/blog/prompt-caching-and-agent-cost/https://pjsofts.github.io/blog/prompt-caching-and-agent-cost/Your agent resends the entire conversation on every single turn, so cost grows quadratically with steps. Prompt caching fixes most of it — but only if your prefix is stable. Here are the four rules, the mistake that silently triples your bill, and why caching and context trimming pull against each other.Tue, 21 Jul 2026 00:00:00 GMTAgentsCostPrompt CachingContext EngineeringProductionPouria Jahandidehhttps://pjsofts.github.io/blog/claude-code-vs-grok-cli-architecture/https://pjsofts.github.io/blog/claude-code-vs-grok-cli-architecture/A source-level comparison of two shipping AI coding agents — which API shape each uses, why Grok speaks OpenAI's wire format on its own servers, and the multi-backend abstraction that lets one client target three APIs.Mon, 20 Jul 2026 00:00:00 GMTClaude CodeGrokAgentsArchitectureAPI DesignPouria Jahandidehhttps://pjsofts.github.io/blog/chat-completions-vs-responses-api/https://pjsofts.github.io/blog/chat-completions-vs-responses-api/The mental-model shift from messages to items, a field-by-field migration map, tool calling with call_id, conversation state, and how the same code runs through the OpenAI SDK, a gateway, and LangChain.Sun, 19 Jul 2026 00:00:00 GMTOpenAIResponses APIAgentsAPI DesignMigrationPouria Jahandidehhttps://pjsofts.github.io/blog/designing-tools-for-ai-agents/https://pjsofts.github.io/blog/designing-tools-for-ai-agents/Better models won't fix badly designed tools. Task-oriented tools over raw APIs, fewer tools over more, designing for the model rather than the human, high-signal returns, and evaluating the tool interface itself.Wed, 15 Jul 2026 00:00:00 GMTAgentsTool DesignEvalsLLMPouria Jahandidehhttps://pjsofts.github.io/blog/what-is-an-agent-loop/https://pjsofts.github.io/blog/what-is-an-agent-loop/Agent frameworks make the loop look complicated. It isn't. Here is the whole thing in Python and TypeScript — the exchange, the stop rule, the step limit, and why writing it yourself changes how you debug everything else.Fri, 10 Jul 2026 00:00:00 GMTAgentsAgent LoopPythonTypeScriptFundamentalsPouria Jahandidehhttps://pjsofts.github.io/blog/context-engineering-for-agents/https://pjsofts.github.io/blog/context-engineering-for-agents/The model has no memory — everything it knows on any turn is something you put in the request. How to assemble a system prompt, budget tokens, load project rules, and compact long sessions without losing what matters.Sun, 05 Jul 2026 00:00:00 GMTContext EngineeringPrompt EngineeringAgentsTokensPouria Jahandidehhttps://pjsofts.github.io/blog/why-your-ai-agent-needs-evals/https://pjsofts.github.io/blog/why-your-ai-agent-needs-evals/You never ship an improvement you haven't measured. Golden datasets, deterministic scorers vs LLM-as-judge, negative test cases, and the data flywheel that turns every user correction into a regression test.Sun, 28 Jun 2026 00:00:00 GMTEvalsTestingAgentsLLM-as-judgePouria Jahandideh