Coding Agent Harness: Plan With Frontier Models, Execute in 90% Less Time

Every serious coding agent - Claude Code, OpenCode, Codex, Cline runs on the same pattern: one model plans, another model executes. Join Varun Krishna, Senior Principal Solutions Engineer at SambaNova Systems on coding agents. Varun has built agentic workflows, model deployment software, and LLM evaluation tooling at SambaNova, with prior experience deploying AI/ML across CRM, e-commerce, healthcare, finance, and more at C3.ai. This session shows you exactly how to build the planning/execution split, cut your token costs by up to 90%, and walk away with your own coding agent harness pointed at SambaNova before the session ends. What You'll Learn 🔹 How the planning and execution split works across today's leading coding agents 🔹 Why offloading execution to a fast, purpose-built inference platform changes the economics of building with AI 🔹 How models handle the execution layer in a real coding agent harness 🔹 How the handoff between planning and execution is structured, step by step 🔹 How to point your own coding agent at a fast inference platform in minutes Why This Matters Most teams building with coding agents are either overpaying or underperforming. Using a frontier model for every step is expensive and slow; using a weak model for everything sacrifices the reasoning quality that makes agents useful. The coding agent harness pattern solves this by matching model capability to task type — frontier models for planning and judgment, fast low-cost models for the high-volume execution work that dominates token usage. As coding agents move from experimental tools into daily workflows, this is one of the most direct ways to cut cost without touching output quality. Who Should Come? Developers integrating coding agents into their workflows, AI engineers exploring cost-efficient inference strategies, and technical leads making infrastructure decisions and evaluating model spend. No prior platform experience required. Bring your questions — we'll leave time for live Q&A.