Technical Computer Science Engineer specializing in autonomous operations and high-integrity AI pipelines.

Project 1: Autonomous KYC & Visa Validation Pipeline

Overview

A mission-critical automation designed to handle high-volume identity verification (Visas/Passports) with a deterministic safety layer to eliminate human error and AI hallucinations.

The Business Problem

Manual verification of identity documents against applicant databases creates a high-latency bottleneck. Reliance on pure AI judgment for legal documents poses a risk of "hallucinations," leading to potential compliance failures and financial loss.

System Architecture

• Ingestion: REST API / Webhooks via Postman.
• Intelligence: Google Gemini 2.5 Flash Vision for OCR extraction of unstructured data.
• Validation Engine: Custom JavaScript implementation of the Levenshtein Distance algorithm to normalize and validate name matches (95% threshold).
• Operational Logic: Conditional routing for "Human-in-the-loop" overrides via Slack API on data discrepancies.

Key Technical Features

• Deterministic Reliability: Uses hard-coded logic as a fail-safe for AI extraction.
• API Resilience: Implements secure header authentication and cross-platform data mapping (n8n —> Airtable —> Slack).
• State Management: Successfully handles multi-stage data passing from trigger to final persistence.

Business Impact (ROI)