PinnedHotaru Komajou·Feb 4From Web Backend to AI Infrastructure — #0: IntroductionThis article series was written to help web backend engineers catch up on the knowledge required for AI and LLM infrastructure.
PinnedHotaru Komajou·Feb 4Now Seeking New Challenges: Senior DevOps & Web Engineer (SoCal)Hello, I’m a Software Engineer based in LA.
Hotaru Komajou·May 17How to Downgrade Google AI Studio from Tier 1 to Free Tier (Without Deleting Your Project)The Problem
Hotaru Komajou·May 10How to Install a Private Discord Bot to Your ServerI always seem to forget the exact steps for this, so I’m putting together this quick guide for my future self. Here is how to create a…
Hotaru Komajou·Feb 20From Web Backend to AI Infrastructure — #5: Distributed System Design for LLM InferenceThe Responsibilities of an LLM Gateway: Building an Intelligent Reverse Proxy
Hotaru Komajou·Feb 18From Web Backend to AI Infrastructure — #4: Advanced Techniques for Inference AccelerationSpeculative Decoding: Drafting with Smaller Models
Hotaru Komajou·Feb 16From Web Backend to AI Infrastructure — #3: Inference Engines — Software-Defined PerformanceThe Modern Inference Engine: vLLM
Hotaru Komajou·Feb 14From Web Backend to AI Infrastructure — #2–2: Memory Optimization for LLM InferenceGPU Optimization: Model Sharding via Parallelism
Hotaru Komajou·Feb 13From Web Backend to AI Infrastructure — #2–1: Memory Optimization for LLM InferencePhysical Constraints Imposed by Data Movement
Hotaru Komajou·Feb 11From Web Backend to AI Infrastructure — #1–2: Understanding Performance Metrics in the LLM EraLong Tails and Percentiles: Improving Core Response Performance