Mihir BrahmaniyaGenAI Intern · AI / Data Science
LinkedIn
All projects
Financial IntelligenceIndustry Project · NoBrokerHood

FinSight AI

Explainable financial intelligence: statement validation, variance, risk and plain-language insights.

My roleDesigned and implemented the system end-to-end, including the AI logic, backend, frontend, data processing and automation workflows.

Explore Try Demo

Problem

Reviewing financial statements for imbalances, suspense balances, unusual movements and risk is manual, and findings are hard to explain to committees and non-accountants.

System pipeline

  1. Financial data
  2. Parse
  3. Classify
  4. Validate
  5. Risk
  6. Insights

Approach

Validation is part of the system, not an afterthought.

Analyses financial statements for imbalances, material movements and risk, and explains every flag with its formula, inputs and threshold. The workflow keeps inputs, decisions and exceptions visible so that a reviewer can understand how the output was reached.

My contribution

  • Statement parsing, detection and account classification
  • Accounting-equation checks and multi-rule analysis
  • Materiality-aware variance, ratios and risk scoring
  • Dashboard, exports and deterministic summary fallback

Try the demo

FinSight AI: interactive demo

24 months of synthetic data → real metrics, variance, anomaly detection, risk flags and insights.

  • Computed
  • Simulated visual
  • Local model
Financial data

Load the sample or a valid CSV.

◇

Load the synthetic dataset or upload a CSV. Results are calculated locally and every flag exposes its rule.

This interactive demonstration is a simplified, synthetic representation inspired by an enterprise project. Production systems, company data and proprietary implementation details are not publicly exposed.

Result

A result that can be inspected.

The portfolio demo calculates revenue, expense, surplus, margin and budget variance, then explains each risk with the observed value, baseline and rule.

Technical architecture

Stage-by-stage system design

  1. 01
    Financial data · Statement ingestion

    XLSX / CSV → Raw rows

  2. 02
    Parse · Statement detection

    Raw rows → Typed rows + statement type

  3. 03
    Classify · Confidence-scored classification

    Typed rows → Account categories + confidence

  4. 04
    Validate · Accounting-equation & rule engine

    Classified accounts → Checks + exceptions

  5. 05
    Risk · Materiality & ratios

    Checks + balances → Risk flags + health score

  6. 06
    Insights · Explainable summary

    Computed findings → Dashboard + report

Tech stack

Tools used in the production system

  • FastAPI
  • Pandas
  • Gemini 2.5 Flash
  • Rule engine
  • SheetJS · jsPDF

The portfolio demo itself is static TypeScript running locally in the browser with synthetic data and no API key.