Harsh Verma https://harshverma.com/ Senior software engineer specializing in retrieval systems, AI evaluation, and production reliability. en Whole articles made a local model accurate, and 15× too slow https://harshverma.com/blog/2026-11-05-local-model-context-budget https://harshverma.com/blog/2026-11-05-local-model-context-budget Thu, 05 Nov 2026 00:00:00 GMT On a local model, every 1,000 tokens of context costs ~2 s. Whole articles reach 0.59 correctness at ~160 s per answer; a context budget gets 0.49 at ~11 s, up from 0.36 with compact context. A frontier model scores 0.65 in 5 s. Latency Evals Local models Qwen3.8 Retrieval RAG Langfuse MCP tool over a durable agent https://harshverma.com/blog/2026-10-19-mcp-tool-durable-agent https://harshverma.com/blog/2026-10-19-mcp-tool-durable-agent Mon, 19 Oct 2026 00:00:00 GMT An MCP tool backed by Temporal workflows: agent runs survive worker crashes, stay within a hard USD cost budget, and work with OpenAI, Claude, or local models. End-to-end OpenTelemetry tracing with token and cost tracking. MCP Budget constraint Temporal Elasticsearch OpenTelemetry Langfuse