https://mohithg.comhttps://mohithg.com/abouthttps://mohithg.com/nowhttps://mohithg.com/useshttps://mohithg.com/work/athenahttps://mohithg.com/work/codebrahmahttps://mohithg.com/work/intuition-aihttps://mohithg.com/work/navyahttps://mohithg.com/work/portfoliopilothttps://mohithg.com/work/ripplinghttps://mohithg.com/writinghttps://mohithg.com/writing/adversarial-evalshttps://mohithg.com/writing/agent-cost-controlhttps://mohithg.com/writing/agent-evalshttps://mohithg.com/writing/agent-failure-modeshttps://mohithg.com/writing/agent-latencyhttps://mohithg.com/writing/agent-loops-are-graphshttps://mohithg.com/writing/agent-observabilityhttps://mohithg.com/writing/agent-state-managementhttps://mohithg.com/writing/agent-tool-permissionshttps://mohithg.com/writing/ai-feature-flagshttps://mohithg.com/writing/ai-features-that-disappearhttps://mohithg.com/writing/ai-product-onboardinghttps://mohithg.com/writing/ai-product-quality-barhttps://mohithg.com/writing/ai-product-roadmaphttps://mohithg.com/writing/ai-product-team-shapehttps://mohithg.com/writing/ai-product-trusthttps://mohithg.com/writing/ai-product-versioninghttps://mohithg.com/writing/build-substance-then-surfacehttps://mohithg.com/writing/continuous-vs-deep-vs-shadow-evalshttps://mohithg.com/writing/cost-attribution-llmhttps://mohithg.com/writing/cost-of-contexthttps://mohithg.com/writing/economics-of-llm-in-productionhttps://mohithg.com/writing/eval-datasets-that-hold-uphttps://mohithg.com/writing/eval-drifthttps://mohithg.com/writing/eval-rubrics-are-the-workhttps://mohithg.com/writing/few-shot-designhttps://mohithg.com/writing/from-demo-to-productionhttps://mohithg.com/writing/hidden-cost-of-evalshttps://mohithg.com/writing/human-in-the-loop-evalshttps://mohithg.com/writing/llm-as-judge-what-workshttps://mohithg.com/writing/llm-batch-vs-realtimehttps://mohithg.com/writing/llm-build-vs-buyhttps://mohithg.com/writing/llm-eval-bench-actually-needshttps://mohithg.com/writing/llm-pricing-tier-designhttps://mohithg.com/writing/llm-rate-limitshttps://mohithg.com/writing/llm-unit-economicshttps://mohithg.com/writing/measuring-ai-product-successhttps://mohithg.com/writing/minimum-viable-eval-benchhttps://mohithg.com/writing/model-routing-costhttps://mohithg.com/writing/multi-agent-vs-single-agenthttps://mohithg.com/writing/optimizing-llm-spend-late-stagehttps://mohithg.com/writing/personas-useful-or-theatrehttps://mohithg.com/writing/prompt-caching-economicshttps://mohithg.com/writing/prompt-versioning-practicalhttps://mohithg.com/writing/prompts-as-type-signatureshttps://mohithg.com/writing/schema-first-promptshttps://mohithg.com/writing/system-prompts-that-age-wellhttps://mohithg.com/writing/system-vs-user-vs-developerhttps://mohithg.com/writing/the-ai-vocabulary-problemhttps://mohithg.com/writing/tool-design-for-agentshttps://mohithg.com/writing/trace-everything-debugginghttps://mohithg.com/writing/when-to-ship-a-prompt-changehttps://mohithg.com/writing/when-to-use-agents