Context
Modernisation of a financial chatbot ecosystem for investment and holdings monitoring.
Problem
Bring financial data, monitoring notes and user documents together in a structured SWOT analysis while preserving the expected format and a fluid real-time experience.
My role
I develop and evolve the SVI agent across the Python services and Angular interface.
- Ingest PDF and DOCX files and maintain an extraction agent for PDF, image, DOCX and PPTX inputs.
- Generate a SWOT with summary, sources and recommendations from financial and documentary context.
- Stream tokens, execution steps, tool calls and results through SSE.
- Export the analysis to PowerPoint and document the transition between Java, Python and MCP architectures.
Delivered result
An end-to-end conversational workflow that enriches LLM analysis with financial and documentary context and delivers reusable SWOT presentations.
Limits
No public adoption, quality or productivity metric was provided. The result therefore describes delivered capabilities, not an inferred business impact.
TOOLS AND FORMATS
- LLM agents
- Python
- Angular
- SSE
- MCP
- Financial analysis
- PDF · DOCX · PPTX
- PowerPoint