Back to portfolio

Engineering work.

Selected work across search, product, shared infrastructure, identity, operations and payments.

Four more engineering examples

PostgreSQL, Redis
& consistency boundaries

Redis pre-checks and more precise cache keys.PostgreSQL / Redis

The problem

A cache entry, relational record and search document represent different versions of the same business fact.

My contribution

I added Redis pre-checks to variant queries, refined manufacturer cache keys and invalidation after successful writes, and worked on targeted relational projections.

The engineering decision

Keep the source of truth explicit. Invalidate after a successful write, and preserve query semantics while reducing unnecessary reads.

Cache keys / Invalidation / Read projections

Event processing
& background jobs

I designed shared event-handling behavior for .NET services..NET / MediatR / Hangfire

The problem

Services needed consistent ways to handle events and failures across multiple operations.

My contribution

In 2024, I designed shared MediatR infrastructure with selectable failure-handling policies. I also developed background media jobs and resolved scheduled-job issues.

The engineering decision

I moved event-handling rules into shared infrastructure so each service could select the behavior it needed.

MediatR / Handler policies / Background processing

Service diagnostics
& operational tooling

Timing and traceability across service workflows.Docker / Kubernetes / ELK

The problem

A failed request may start in one service and become visible in another. Understanding the runtime matters as much as reading the code.

My contribution

I diagnose service failures through Elasticsearch, application logs, background jobs and container state, working with delivery and operational tooling.

The engineering decision

Use logs, timing and service state together to narrow down a failure before changing behaviour.

Container environments / Logs / Cross-service troubleshooting

Payment integrations
& reconciliation

Provider integrations, refunds and reconciliation safeguards..NET / Kafka / ERP

The problem

Payment providers, order records and ERP documents can disagree after a timeout or a partial refund.

My contribution

I built payment provider integrations, cancellation and refund flows, event-driven state handling and checks around downstream accounting operations.

The engineering decision

Treat an uncertain provider result explicitly. Reconcile before repeating a financial operation.

Payment providers / Cancellation and refunds / ERP integration