Case studies
The problem, the architecture, the decisions and what I learned, for each project.
Merchant Dashboard Agent
An embedded AI assistant that turns plain-language questions into live dashboard views for a multi-tenant payments platform.
- Role
- Lead and sole engineer
- Timeframe
- About 4 months, 2026 (1 month core build, 3 months of upgrades and testing)
- Stack
- Python, Django, Anthropic Claude, SSE, ApexCharts, PostgreSQL, Playwright
Multi-Tenant Access Control for SaaS
Custom roles, section-wise permissions, cross-tenant access and tenant isolation for a schema-per-tenant B2B payments SaaS.
- Role
- Lead engineer for access control and tenant isolation
- Timeframe
- About 3.5 months, 2026
- Stack
- Python, Django, django-tenants, PostgreSQL (schema per tenant), Next.js, Playwright
Fuxx (Liquva) SEPA Payment Integration
A SEPA Direct Debit provider integrated by API and by CSV batch, with signed webhooks, polling and reconciliation for a multi-tenant payments SaaS.
- Role
- Lead engineer for the Fuxx integration
- Timeframe
- About 2 months, 2026
- Stack
- Python, Django, PostgreSQL, Celery, Node.js/TypeScript (provider mock), Playwright
Customer Communications, Part 1: Email
Event-driven transactional email for a multi-tenant SEPA Direct Debit SaaS, sent through each merchant's own mail server.
- Role
- Lead and sole engineer
- Timeframe
- About 3 months, 2026
- Stack
- Python, Django, Celery, Redis, PostgreSQL, Jinja2, Datadog, Playwright
Customer Communications, Part 2: Physical Letters
Printed and posted letters as a second channel on the email platform, through the Pingen print-and-mail API, with per-merchant cost and delivery tracking.
- Role
- Lead and sole engineer
- Timeframe
- About 1 month, 2026
- Stack
- Python, Django, Celery, PostgreSQL, Gotenberg (HTML to PDF), Pingen API, Next.js