https://muhammadamal.my.id/tags/agents/ Recent content in Agents on Hi, I'm Muhammad Amal Hugo en-us Fri, 24 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/evaluating-llm-agents/ Fri, 24 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/evaluating-llm-agents/ A practical agent evaluation system with deterministic checks, LLM-as-judge rubrics, and the regression discipline that survives model upgrades. https://muhammadamal.my.id/blog/cost-control-for-llm-agents/ Wed, 22 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/cost-control-for-llm-agents/ How to keep agent bills under control with hard budgets, per-step caps, model routing, and the prompt caching feature Anthropic just shipped. https://muhammadamal.my.id/blog/guardrails-for-llm-agents-2024/ Mon, 20 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/guardrails-for-llm-agents-2024/ Where Llama Guard, Rebuff, and NeMo each earn their place in an agent guardrail stack, and the layered approach that keeps the false positive rate bearable. https://muhammadamal.my.id/blog/memory-for-ai-agents/ Fri, 17 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/memory-for-ai-agents/ How to give agents memory without overstuffing the context window or building a vector database mausoleum, with concrete patterns and code. https://muhammadamal.my.id/blog/react-reflexion-planner-executor-agent-loops/ Wed, 15 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/react-reflexion-planner-executor-agent-loops/ Three agent loop patterns, when each is the right shape, and the code to build them with LangChain 0.1 and OpenAI’s function calling. https://muhammadamal.my.id/blog/multi-agent-conversations-with-autogen/ Mon, 13 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/multi-agent-conversations-with-autogen/ When AutoGen’s multi-agent model earns its keep, when it does not, and the patterns that prevent infinite loops and budget blowups. https://muhammadamal.my.id/blog/designing-tools-for-llm-agents/ Fri, 10 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/designing-tools-for-llm-agents/ The tool layer is where most agents quietly fail. This is the schema, validation, and error-design playbook that keeps yours working. https://muhammadamal.my.id/blog/production-agents-langgraph-state-machines/ Wed, 08 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/production-agents-langgraph-state-machines/ A practical walkthrough of building agents that survive production with LangGraph state machines, checkpointing, and explicit edges. https://muhammadamal.my.id/blog/agentic-ai-landscape-may-2024/ Mon, 06 May 2024 09:00:00 +0700 https://muhammadamal.my.id/blog/agentic-ai-landscape-may-2024/ Comparing the three agent frameworks that matter right now, with code, and the architectural decisions that follow from each.