https://miasyster.github.io/en/tags/ai-agent/Recent content in AI Agent on AI + Quant EngineeringHugoenFri, 08 May 2026 00:00:00 +0000https://miasyster.github.io/en/posts/cross-model-review-is-architecture/Fri, 08 May 2026 00:00:00 +0000https://miasyster.github.io/en/posts/cross-model-review-is-architecture/Single-LLM self-reflection has a structural blind spot — the model tends to confirm its own output. QuantGPT enforces a hard rule in factor-mine SKILL Phase 0.5: Claude must consult DeepSeek before designing a new factor family. Not a suggestion, a hard rule. This isn’t redundancy — it’s the antidote to structural bias.https://miasyster.github.io/en/posts/agent-native-architecture/Fri, 01 May 2026 00:00:00 +0000https://miasyster.github.io/en/posts/agent-native-architecture/When the system operator changes from a human to an LLM Agent, design principles need fundamental rethinking. Humans need GUIs and documentation. Agents need semantically clear tools and constraints that throw errors.https://miasyster.github.io/en/posts/harness-is-governance/Fri, 01 May 2026 00:00:00 +0000https://miasyster.github.io/en/posts/harness-is-governance/The mainstream approach to Agent governance is writing rules in prompts. But LLMs can ignore prompts. Real governance lives outside the Agent, in the Harness layer — tools define what’s possible, errors define what’s forbidden, code defines where the boundaries are.https://miasyster.github.io/en/posts/why-agent-infra-not-agent/Fri, 01 May 2026 00:00:00 +0000https://miasyster.github.io/en/posts/why-agent-infra-not-agent/While everyone is building Agent frameworks, I chose a different path: build infrastructure for Agents, not the Agent itself. Not because I can’t build Agents, but because the Agent layer is a consumable — infrastructure is an asset.https://miasyster.github.io/en/posts/skill-orchestration-over-agent-loops/Thu, 30 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/skill-orchestration-over-agent-loops/Current AI Agent frameworks obsess over building complex loop chains: Planner → Executor → Reflector → Re-planner. I chose the opposite: tools are stateless pure functions, and the LLM decides the call sequence itself. Not because loop chains aren’t cool — but because they put decision authority in the wrong place.https://miasyster.github.io/en/posts/sandbox-defense-in-depth/Fri, 03 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/sandbox-defense-in-depth/AI-generated code must be executed — otherwise it’s just text. But execution means risk. I didn’t choose container isolation or RestrictedPython. Instead I designed a three-layer defense: reject dangerous structures at compile time via AST, replace the entire builtins at runtime, and enforce OS-level resource limits as a backstop. Each layer handles a different class of risk. Overlapping but not redundant.https://miasyster.github.io/en/posts/why-i-built-my-own-fsm-engine/Fri, 03 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/why-i-built-my-own-fsm-engine/LangChain, LangGraph, CrewAI, PydanticAI — no shortage of AI orchestration frameworks. I evaluated all of them and built my own. Not NIH syndrome. When you need failure-mode-driven mutation strategies, phase-aware multi-model routing with different temperatures, and adaptive evolution based on trajectory analysis, the abstraction layers of general-purpose frameworks become obstacles to route around.https://miasyster.github.io/en/posts/ai-as-operator/Thu, 02 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/ai-as-operator/Letting AI drive research workflows doesn’t mean letting AI decide how the system runs. I made a key separation: AI is just the operator, the execution engine is the law. This decision came from a failure.https://miasyster.github.io/en/posts/design-for-the-endgame/Thu, 02 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/design-for-the-endgame/Most systems are designed to run first, then audited as an afterthought. I inverted the order — first define what questions the system must answer when things go wrong, then work backwards to what each layer must record. This inversion reshaped the entire architecture.https://miasyster.github.io/en/posts/mcp-semantic-gap/Thu, 02 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/mcp-semantic-gap/MCP’s JSON-RPC transport works fine. The real problem: natural language rules have no code-level enforcement — the LLM can completely ignore your instructions. I designed the Intent Validator pattern to close this gap.https://miasyster.github.io/en/posts/why-not-multi-agent/Thu, 02 Apr 2026 00:00:00 +0000https://miasyster.github.io/en/posts/why-not-multi-agent/Multi-agent is the hot paradigm in AI engineering. I chose single agent + state machine for my AI-driven quant research system. Not because multi-agent is too hard, but because the problem structure doesn’t match.