It turns local data into governed analysis, models, strategies, and audit-ready deliverables.
One workbench for local risk analysis and development—from data to strategy and reports.
From a request to a reviewable result
MARVIS is a local-first, governed credit-risk Agent platform—not a chatbot wrapped around a collection of scripts.
Describe the business outcome in natural language. MARVIS asks for missing files and definitions, builds a reviewable plan, pauses at responsibility gates, runs deterministic tools, and returns real datasets, evidence, models, strategy code, and reports.
flowchart LR
A["Describe the risk goal"] --> B["Agent clarifies inputs and definitions"]
B --> C["Validated workflow plan"]
C --> D{"Human confirmation<br/>where required"}
D --> E["Deterministic tools execute"]
E --> F["Evidence, artifacts, and reports"]
F --> G["Review, adopt, and iterate"]
What the current V2 delivers
- A complete seven-step strategy-development workflow: current and historical evidence, governed dual-population samples, univariate and model evidence, trees, Cross, scorecards, Voting, Strategy Pools, impact measurement, validation, code delivery, and four-format review reports.
- A governed data-to-model workflow: ingest and join files, analyze and engineer features, train and compare multiple recipes, export PMML, score data, generate model reports, and hand the selected model and supporting evidence directly into model validation. PMML export is available for supported recipes.
- Conversational risk analysis: MARVIS first asks what to analyze and which fields, units, dates, scenarios, and assumptions apply. It then runs the selected VTG-terminal/annualized-bad-rate or profitability calculation and delivers an audited Excel report. Standard Vintage and roll-rate are separate governed workflows with structured evidence and artifacts.
Why risk teams use MARVIS
| Work in business language Start with the decision you need, not a hand-built chain of scripts and notebooks. |
Trust the numbers KS, AUC, PSI, bad rate, approval rate, profit, and impact are calculated by deterministic platform code—not guessed by an LLM. |
Keep data close Files, task state, evidence, and outputs stay in a controlled local workspace by default. |
Retain human responsibility High-impact actions pause for confirmation, and key governed results carry lineage and audit evidence. |
Agent mode and the Manual Workbench share the same validated workflows, tools, schemas, and deterministic calculation kernels.
No comments yet.