Portfolio
Projects
Agentic AI workflows with LLMs, APIs, n8n and Zapier.
Requirements to Jira Stories, QA Tickets and Test Script (n8n + Claude)
An n8n workflow that turns an FSD or Change Request (Word/PDF) and wireframes into Jira user stories with acceptance criteria, linked QA tickets and a bank-style Excel test script. It writes missing acceptance criteria and never recreates existing stories.
RAG FAQ Chatbot with n8n, Claude & Qdrant
An end-to-end customer-support chatbot built in n8n that answers only from FAQ documents, cites its sources, refuses to guess, and turns every unanswered question into a to-do item for the FAQ owner. Ships with an automated evaluation suite and quality gates.
AI Test-Case Generator for Jira (n8n + Claude)
An n8n workflow that turns Jira user stories into traceable manual test cases (positive, negative, boundary, security), verifies acceptance-criteria coverage with a code quality gate, posts the results back to Jira, and asks the product owner questions instead of inventing requirements.

AI Data Quality Monitor (n8n + Claude + PostgreSQL)
An n8n workflow that runs SQL data quality rules on a mobile money database, manages incidents, and has Claude explain each new failure, with a code guardrail before anything reaches Jira. Caught 16 of 16 planted defects with 0 false alarms, for about $0.15 a run.
Regulatory Change Tracker for AI & Data Governance (n8n + Claude)
An n8n workflow that watches regulators in any market (EU, US, UK, Australia, Zimbabwe) for new and amended AI and data governance publications. Haiku triages, Opus writes briefs with quoted dates, obligations and affected controls, code verifies every quote, and a person approves before Jira tasks are created.

KYC Document Review with AI Extraction and Human Decision (n8n + Claude)
Reads identity documents and proofs of address from photos, HEIC, scans, PDFs and Word files; handles password-protected files without the password touching the workflow; Claude reads, code checks passport MRZ check digits, expiry, age and cross-document matches, and a person decides. 48 of 48 fields read exactly, about $0.05 per application.