Selected experience
Products, infrastructure, and engineering backgroundEvidence over labels
Si-Jam is a small engineering company. The value is direct technical involvement: understanding the existing system, finding the first failing layer, making controlled changes, and leaving behind tests and documentation rather than a new black box.
End-to-end published Android product
GeldPlan
Personal cashflow planning and expense tracking
GeldPlan is an Android application developed and published by SI-DZHEM, TOO under the Si-Jam name. It treats money as events in time so a user can record actual and planned transactions and understand the operational balance ahead. Product concept, engineering, visual design, graphics, publication, and production support were handled within Si-Jam.
Product engineering
Kotlin, Jetpack Compose, Room persistence, dense operational UI, localization, and iterative product framing.
Visual design
The application interface, icons, and other visual assets were created by the Si-Jam designer alongside product development.
Data safety
Explicit schema migrations, backup and restore, legacy-data handling, and careful currency semantics without silently inventing financial facts.
Offline-first behavior
Core records and planning remain useful locally; network-backed rates and platform services are supporting capabilities rather than the data model.
Release operations
Google Play delivery, privacy-policy synchronization, consent and store requirements, release checks, and production issue diagnosis.
Operational platform
Si-Jam production infrastructure
The company site and internal services run on a heterogeneous three-node K3s control-plane. The platform is maintained as real production infrastructure rather than a slide-deck architecture: deployments, storage, certificates, monitoring, incidents, node failures, and recovery are handled against live state.
High availability
Three control-plane/etcd nodes, kube-vip service addresses, health verification, and deliberate failure-domain checks.
Delivery
GitLab CI, Kaniko and multi-architecture OCI images, Helm releases, immutable tags, and staged production rollout.
State and recovery
Longhorn-backed persistent storage, certificate management, backup-aware operations, and evidence-led recovery after host or power failures.
Observability
Metrics, logs, alerts, service-level checks, and runbooks that connect symptoms to the first failing component.
Professional background
From networks and databases to embedded systems and DevOps
2002–2011: networks and operations
Linux/FreeBSD servers, routing and branch connectivity, monitoring, VoIP, a MySQL cluster, and a Perl/MySQL/RADIUS billing system. Technical leadership contributed to regional ISP operations serving more than 10,000 subscribers.
2011–2019: ERP and data
Legacy ERP support, production material-flow analysis, database architecture, a replacement backend, frontend integration, and procurement and warehouse planning data.
Since 2019: embedded engineering
Device development, legacy desktop and embedded support, reusable hardware-level libraries, team leadership, documented development processes, and software and hardware tests.
Current delivery work
CI/CD, containers, Kubernetes, hybrid and offline deployment, application bootstrap, production diagnosis, Android development, and reliability-focused documentation.
Much of the embedded and infrastructure work is proprietary. Relevant details can be discussed at an appropriate level during an initial conversation without exposing a customer's code, data, or internal architecture.