I like the part of software where a wrong number or a lost write is a real bug. Most of my work is in Go: risk systems that can replay any past state, collaboration servers that never lose an edit, and network automation that catches drift before it breaks anything.
|
BACKEND Go services and REST APIs on PostgreSQL and Redis: multi-tenant isolation with row-level security, append-only data models, background jobs and usage-based billing. |
REAL-TIME WebSocket services and CRDT-based collaboration (Yjs) with durable persistence, horizontal scaling and conflict-free concurrent editing. |
|
FINTECH Point-in-time market data, exact-decimal valuation, reconciliation, stress testing and ledgers where every number can be traced back to its source. |
INFRASTRUCTURE Reproducible network labs and intended-state automation: OSPF/BGP/IPsec topologies, NetBox-driven Ansible deploys, rollback and drift detection. |
|
FINTECH · RISK Personal risk OS that joins macro data, portfolio exposure, stress tests and decisions on one point-in-time timeline, with backup/restore and a fail-closed release gate.
|
BACKEND · SAAS Multi-tenant project management with usage-based billing, built as a Go modular monolith with Postgres row-level security and Stripe metered billing.
|
|
REAL-TIME · DISTRIBUTED Horizontally scalable Yjs collaboration server: WebSocket sync with CRDT persistence to Postgres, Redis and S3, plus auth, metrics and audit logging.
|
INFRASTRUCTURE · AUTOMATION Intended-state network automation: NetBox YAML drives gated Ansible deploys with rollback and drift detection.
|



