Build an AI-Powered Financial Market Intelligence System
The rapid shifts in financial valuations highlight the need for real-time, sophisticated market intelligence. This challenge tasks you with building an autonomous agent system using a graph-based workflow to monitor financial news, analyze company valuations, predict market impacts, and identify investment opportunities or risks. Your system will leverage Gemini 2.5 Pro's advanced reasoning capabilities, particularly its 'Deep Think' mode for complex financial modeling, combined with OpenAI o3 for instant summarization. The core orchestration will be built with LangGraph, enabling dynamic, stateful decision-making workflows. It will also integrate with external financial data APIs via an MCP (Multi-Agent Communication Protocol) server for real-time data ingestion, and employ RAG (Retrieval Augmented Generation) over a corpus of financial reports and news archives.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop a LangGraph-orchestrated multi-agent system utilizing Gemini 2.5 Pro and OpenAI o3 for advanced financial market intelligence with real-time data integration.
How work is evaluated
The evaluation will assess the agent system's ability to accurately and comprehensively analyze financial market data, generate insights, and integrate tools effectively within a LangGraph workflow.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Implement extended thinking with adaptive reasoning budgets using Gemini 2.5 Pro.
Generate actionable financial insights and risk assessments.
Design a graph-based agent workflow for complex financial analysis.
Integrate real-time financial data via MCP-enabled tools.
Develop a hybrid reasoning system for varied analytical tasks.
Reference links and supporting material
Sample data for 2 tasks
Challenge setup
The evaluation will assess the agent system's ability to accurately and comprehensively analyze financial market data, generate insights, and integrate tools effectively within a LangGraph workflow.
Challenge input
{'news_articles': [{'title': '...', 'content': '...', 'source': '...'}], 'target_companies': ['Revolut', 'Tempo']}
Agent execution
No agent tool is listed for this challenge.
Evaluated output
{'company_analysis': [{'company': 'Revolut', 'sentiment': 'positive/negative/neutral', 'key_events': ['...'], 'potential_impact': '...'}], 'overall...
- Verify the LangGraph workflow executes without errors and transitions between states correctly.
- Confirm that MCP-enabled tools are successfully called and return data, as evidenced by logs or output.
- Check if both quick summarization (OpenAI o3) and deep analysis (Gemini 2.5 Pro Deep Think) components are...
- Analysis Accuracy target: 0.9
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Verify the LangGraph workflow executes without errors and transitions between states correctly.
- Confirm that MCP-enabled tools are successfully called and return data, as evidenced by logs or output.
- Check if both quick summarization (OpenAI o3) and deep analysis (Gemini 2.5 Pro Deep Think) components are utilized as intended.
Metrics
- Analysis_Accuracy · Target 0.9
- Insight_Depth · Target 4
- Latency_for_Instant_Tasks · Target 2
Workflow tasks
- Task 1
Input: {'news_articles': [{'title': '...', 'content': '...', 'source': '...'}], 'target_companies': ['Revolut', 'Tempo']}
Output: {'company_analysis': [{'company': 'Revolut', 'sentiment': 'positive/negative/neutral', 'key_events': ['...'], 'potential_impact': '...'}], 'overall_market_summary': '...'}
Execution harness: Python
- Task 2
Input: {'company_symbol': 'GOOGL', 'risk_factors_prompt': 'Assess geopolitical risks and market competition.'}
Output: {'financial_data': {'current_price': 180.50, 'analyst_rating': 'Strong Buy', 'pe_ratio': 28.5}, 'risk_assessment': {'geopolitical_risk': 'low/medium/high', 'competition_risk': 'low/medium/high', 'overall_assessment_summary': '...'}}
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.