---
title: "Create a LangChain Deep Agents Harness Profile for NVIDIA Nemotron 3 Ultra to Improve Performance"
description: "The tutorial shows how to create and validate a LangChain Deep Agents harness profile for NVIDIA Nemotron 3 Ultra using an evaluation benchmark rather than fine-tuning the model. It starts with a baseline, analyzes failed trajectories, and iterates on prompts, exclusions, or added middleware while rerunning tests to detect regressions. One failure occurred when read_file returned the first page of a large file without signaling that more content existed, so the model answered before paginating to the line containing opal-fox-91. The proposed ReadFileContinuationNoticeMiddleware appends a continuation notice when a full page is returned, raising the reported benchmark score from 94/127 to 96/127 and resolving all three read_file tests. The source also describes automating this propose, verify, and retain-or-reject loop, with safeguards against overfitting and regressions."
---

# Create a LangChain Deep Agents Harness Profile for NVIDIA Nemotron 3 Ultra to Improve Performance

[NVIDIA](https://yomu.fyi/company/nvidia-developer-blog) · Sean Lopp · Jul 8, 2026

**Type:** Tutorial

## Summary

The tutorial shows how to create and validate a LangChain Deep Agents harness profile for NVIDIA Nemotron 3 Ultra using an evaluation benchmark rather than fine-tuning the model. It starts with a baseline, analyzes failed trajectories, and iterates on prompts, exclusions, or added middleware while rerunning tests to detect regressions. One failure occurred when read\_file returned the first page of a large file without signaling that more content existed, so the model answered before paginating to the line containing opal-fox-91. The proposed ReadFileContinuationNoticeMiddleware appends a continuation notice when a full page is returned, raising the reported benchmark score from 94/127 to 96/127 and resolving all three read\_file tests. The source also describes automating this propose, verify, and retain-or-reject loop, with safeguards against overfitting and regressions.

## Context

Agentic systems face a trade-off between accuracy and cost, while fine-tuning smaller open models requires specialized expertise and training and hosting hardware. The tutorial presents harness-specific evaluation and customization as a way to improve model performance through agent-harness changes using existing NVIDIA Nemotron 3 Ultra endpoints.

## Approach / What changed

Create a LangChain Deep Agents HarnessProfile for Nemotron 3 Ultra, establish a benchmark baseline, inspect failed trajectories, and modify prompts, exclusions, or middleware. The example adds ReadFileContinuationNoticeMiddleware, which detects a full read\_file page and appends instructions to continue with offset pagination. Changes are rerun against the failing test and the full benchmark, with repeated verification and optional held-out validation to reduce overfitting.

## Takeaways

- The read\_file failure occurred because a full first page contained no signal that the file continued, causing Nemotron 3 Ultra to answer before reaching the target line.
- ReadFileContinuationNoticeMiddleware appends a notice when the returned line count reaches the per-read limit, directing the model to call read\_file again with an increased offset.
- The reported profile change improved the evaluation score from 94/127 to 96/127 and resolved all three failing read\_file tests; repeated runs matter because the benchmark is stochastic.

**Tags:** [LLMs](https://yomu.fyi/topic/llm), [Open Source](https://yomu.fyi/topic/open-source), [Performance](https://yomu.fyi/topic/performance), [Python](https://yomu.fyi/topic/python), [Testing](https://yomu.fyi/topic/testing)

- Source: [NVIDIA](https://developer.nvidia.com/blog/create-a-langchain-deep-agents-harness-profile-for-nvidia-nemotron-3-ultra-to-improve-performance)
- Source URL: https://developer.nvidia.com/blog/create-a-langchain-deep-agents-harness-profile-for-nvidia-nemotron-3-ultra-to-improve-performance
- Ingested by Yomu: 2026-08-27T14:59:43.588Z

[Read original post](https://developer.nvidia.com/blog/create-a-langchain-deep-agents-harness-profile-for-nvidia-nemotron-3-ultra-to-improve-performance)