Challenge

Extended Reasoning Orchestration with Claude Agents SDK and Milvus

Design an enterprise planning architecture using Claude Agents SDK to manage complex document synthesis, inspired by the surge in public interest in classic literature reissues. By utilizing GPT-5.4 Pro as a reasoning engine and Milvus for high-performance retrieval of structured event data, you will build an agent capable of performing multi-step analysis on lengthy texts, ensuring reasoning is persistent and globally indexed across long-horizon projects.

Machine LearningHosted by Vera
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Orchestrate advanced reasoning agents with Claude Agents SDK and Milvus to analyze and synthesize large-scale document collections.

Delivery guide

How work is evaluated

Evaluation

Multi-step reasoning accuracy and retrieval performance.

Datasets

Shared data for this challenge

Review public datasets and any private uploads tied to your build.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

SynthesisCompleteness

Ensure no key plot points missed

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

RetrievalPrecision

Milvus recall score • target: 0.95 • range: 0.85-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Master long-context reasoning using Claude Agents SDK for complex synthesis

  • Implement persistent memory buffers using Milvus for high-speed retrieval

  • Orchestrate hybrid reasoning cycles combining GPT-5.4 Pro and Claude Sonnet 4.6.6

  • Deploy optimized inference pipelines using Novita AI

  • Build self-correcting agents that verify facts against indexed datasets

  • Design advanced observability dashboards for multi-model decision traces

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

How this agent runs

Multi-step reasoning accuracy and retrieval performance.

Challenge input

Complex document text

Milvus

Vector DB

Zed

High-performance code editor

GPT-5

Policy Serving in the agent workflow.

Evaluated output

Structured synthesis report

Checks for
  • Ensure no key plot points missed
  • Milvus recall score • target: 0.95 • range: 0.85-1
Proof of success
  • RetrievalPrecision target: 0.95
  • 1 public reference case
Runtime evidence
  • Docker execution harness
View technical recipe

Configured tools

Action Space
  • Milvus · Required
  • Zed · Optional
Policy Serving
  • GPT-5 · Optional

Evaluation contract

  • SynthesisCompleteness · Weight 1
  • RetrievalPrecision · Weight 1

Recipe state

This is a preview. The configuration can change before the evaluation recipe is locked.

Frequently Asked Questions about Extended Reasoning Orchestration with Claude Agents SDK and Milvus