https://mhassan.dev/categories/ai-systems/Recent content in AI Systems on Muhammad Hassan RazaHugoen-USFri, 25 Sep 2026 10:00:00 +0500https://mhassan.dev/blog/langgraph-multi-agent-middleware/Wed, 08 Apr 2026 10:00:00 +0500https://mhassan.dev/blog/langgraph-multi-agent-middleware/Obelisk runs more than fifteen specialist agents. The agents were the easy part. Model choice, context, cost, retries and approvals all ended up in a middleware layer, and that layer is what made it work.https://mhassan.dev/blog/fraud-scoring-legal-intake/Wed, 08 Apr 2026 10:00:00 +0500https://mhassan.dev/blog/fraud-scoring-legal-intake/Coached callers, invented doctors and borrowed photos. How RISQ scores intake calls for mass tort firms, and why the most useful score measures the interviewer, not the caller.https://mhassan.dev/blog/langchain-production/Sun, 01 Feb 2026 10:00:00 +0500https://mhassan.dev/blog/langchain-production/Where LangChain earns its place, where it gets in the way, and the handful of habits that kept our AI features from failing quietly: timeouts, validation, real persistence, tracing and pinned versions.