NSNeeloppher Syedinneeloppher.hashnode.dev·4d ago · 8 min readMoE Cost Analyzer: Benchmark Dense vs Mixture-of-Experts LLMs on Your Own Prompts with Real API CallsEvery team running LLMs in production eventually asks the same question: should we switch to a Mixture-of-Experts model? The MoE architecture activates fewer parameters per token, which means cheaper 11K
NSNeeloppher Syedinneeloppher.hashnode.dev·Jul 17 · 5 min readInfra-Sizer: A CLI That Recommends the Right GPU Infrastructure for Any LLM Deployment"What hardware do I need to serve this model?" is the first question every team asks before deployment. The answer requires juggling model size, throughput benchmarks, AWS pricing, batching math, and 11K
NSNeeloppher Syedinneeloppher.hashnode.dev·Jul 17 · 8 min readNeuralUCB Router: An OpenAI-Compatible API Proxy That Routes LLM Requests Using a Multi-Armed Bandit AlgorithmEvery team using LLMs in production hits the same wall: you don't need GPT-4 for every request, but you don't know in advance which ones actually need it. A user asking "what is 2+2?" and a user askin00
NSNeeloppher Syedinneeloppher.hashnode.dev·Jul 15 · 11 min readAI Cost Optimizer CLI: Real-Time AI Cost Tracking, Model Routing, and Shadow AI Detection from the TerminalEngineering teams running AI at scale eventually hit the same wall: no one knows exactly how much each model call costs, which GPU is idle, whether a cheaper model would have done the job just as well00
NSNeeloppher Syedinneeloppher.hashnode.dev·Jul 10 · 15 min readDataset Factory: A Production-Grade Benchmark Dataset Factory for AI Agent EvaluationEvaluating AI agents requires benchmark datasets that are high-quality, diverse, balanced, and free of duplicates. Building those datasets by hand is slow, inconsistent, and hard to reproduce. The Mer00
