My Role
AI product and full-stack engineer
All Projects
Live prototypeFlagship case study
FinSight
AI Financial & Document Intelligence
Financial records, identity documents, and card activity require different analysis workflows, but users need one explainable view of the resulting risk and insight.
Interactive preview
Full demoWhat I Built
AI-powered financial intelligence platform that analyzes complex documents, extracts structured evidence, and adds a credit-card intelligence workspace for comparing rewards, fees, spending profiles, and first-year value. Built with FastAPI, React, AWS Bedrock, LandingAI ADE, and explainable recommendation logic.
Evidence
Working interface, documented system behavior, and implementation-level decisions.
Technical Architecture
From system input to explainable output.
The control gate is shown as a first-class stage, not an afterthought added around the workflow.
Technical Decisions
- Built an end-to-end financial document pipeline on FastAPI and React that ingests statements and identity documents, extracts structured fields, and produces an audit-ready result set.
- Integrated LandingAI ADE for layout-aware extraction so tables and key-value regions survive the parse instead of collapsing into raw text.
- Orchestrated agentic reasoning over the extracted fields with AWS Bedrock models, splitting extraction, cross-document reconciliation, and compliance judgement into separate steps.
- Implemented traceability so every generated conclusion links back to the document, page, and region it came from, which is the difference between a demo and something an auditor accepts.
Security Considerations
- Built an end-to-end financial document pipeline on FastAPI and React that ingests statements and identity documents, extracts structured fields, and produces an audit-ready result set.
- Implemented traceability so every generated conclusion links back to the document, page, and region it came from, which is the difference between a demo and something an auditor accepts.
Outcome & Evidence
- Automated KYC and AML checks including identity consistency, sanctions-style name matching, and transaction pattern flags, each with the source field cited.
- Implemented traceability so every generated conclusion links back to the document, page, and region it came from, which is the difference between a demo and something an auditor accepts.
FastAPIReactAWS BedrockLandingAI ADEDockerFinTechCredit IntelligenceRecommendation Systems
Working product
Try the interactive demo.
The product experience is part of this case study. Explore it here, reset its state, or switch viewport sizes without leaving the project page.
finsight.zainkhalilkhan.com
This is the actual deployed application, embedded live from v0-fin-sight11-av.vercel.app. If it doesn't load below (some browsers block third-party embeds), open it directly.
Loading live application...
Client-side sandbox. State is in memory and nothing is sent to a server.
Next Case Study