Arkadiusz Szarama

Engineering operations platform spanning workflow orchestration, integrations, and engineering intelligence

• Designed and built the integration platform forming the data backbone of the product, an extensible foundation enabling consistent, reliable connections across 50+ external systems spanning security, incident management, CI/CD, cloud, and observability
• Solved the hard cross-cutting problems at scale: uniform handling of auth diversity, per-provider rate limiting, schema normalization, and failure isolation, so each new integration inherits reliability by default
• Designed and implemented the customer-facing workflow engine, enabling automation of multi-step engineering processes using reusable, authorization-aware building blocks
• Led cross-functional projects end-to-end across product, design, and frontend/backend engineering; authored and reviewed technical specifications that established shared contracts across teams — including the idempotency layer adopted across the Workflows API by other backend engineers
• Extended the Engineering Intelligence layer with multi-source querying, aggregation, and validation for a unified operational view

High-scale, multi-country commerce platform

• Designed and built the entire distributed backend from scratch as founding engineer: 15 microservices covering orders, pricing, inventory, and reporting, with full ownership of every architectural decision
• Established the architectural standards, coding patterns, and onboarding foundations that the team built on as it grew from one to 10+ engineers across backend, QA, and frontend
• Solved distributed consistency under high-concurrency order processing, guaranteeing correctness across service boundaries without a centralized coordinator
• Built the external systems integration layer (ERP, PIM, POS), enabling reliable bidirectional data synchronization across the platform
• Defined service boundaries, data ownership models, and communication patterns that scaled from zero to 300M+ PLN GMV across 13 countries

Real-time telemetry processing at scale

• Re-architected a brittle, data-losing legacy pipeline into a concurrent, event-driven distributed system, achieving a ~20x increase in burst capacity and eliminating data loss under peak load
• Designed a compatibility bridge enabling zero-downtime migration: the new Java system consumed live data from existing C++ and Erlang components in production while the transition was in progress, with full legacy compatibility maintained throughout
• Eliminated data loss by replacing fire-and-forget UDP with guaranteed delivery: persistent queues, acknowledgment-based processing, and automatic retries
• Introduced near real-time processing across ingestion, processing, and reporting layers, supporting ~50K active devices emitting millions of events per day
• Owned the full scope: architecture, technical specification, implementation, and external coordination — including scoping, documentation, and technical sign-off for a sub-project delivered by an outside software agency