09/11/2026 updated
100 % available
Senior AI Software Engineer | MCP , LLM Specialist
WARSZAWA, Poland
Only remote
Bachelor of Arts in Computer ScienceAbout me
Senior software engineer with 11+ years delivering production systems across infrastructure automation, endpoint security, international consulting, and freelance environments, with strong recent focus on C++ and systems-oriented development.
C++ (Programming Language)Embedded CEmbedded SystemsEmbedded Operating SystemsEmbedded SoftwareMessaging Application Programming InterfaceSoftware ArchitectureSoftware Engineering ProcessData StreamingSystems IntegrationWorkflowsBackendEmbedded Coder
• Designed stateful realtime orchestration flows for streaming speech-to-text, LLM decisioning, tool execution, and text-to-speech, keeping interruption handling and live-response latency independent from slower downstream retries and business operations.
• Implemented durable tool workflows for bookings, notifications, forwarding, and other side effects using persistent session state, idempotency keys, explicit timeouts, bounded retries, and event-level audit records to prevent duplicate actions after partial failures.
• Defined guarded escalation and priority-routing logic around LLM decisions, combining explicit rules, required-field checks, confirmations, and safe fallbacks for critical workflows instead of relying on model output alone.
• Built session-oriented backend services that coordinated synchronous user interactions with asynchronous tool calls, persisting workflow state so requests could resume safely after API timeouts, process restarts, or transient infrastructure failures.
• Designed source-of-truth and synchronization patterns for external scheduling and operational systems, preventing competing booking authorities and making downstream calendar views projections of authoritative business state.
• Added end-to-end observability across request, model/tool, and integration paths with correlation IDs, structured logs, latency metrics, outcome events, and replayable failure context, shortening diagnosis of production incidents.
• Expanded automated coverage for interruption, timeout, retry, duplicate-delivery, malformed-response, and partial-completion scenarios; converted recurring production failures into regression tests and explicit recovery contracts.
• Implemented resilient API and event-processing components with explicit state transitions, validation, retries, deduplication, and structured error handling so transient failures did not create inconsistent downstream state.
• Hardened critical workflows with least-privilege service access, auditable action records, controlled fallback paths, and reviewable business rules around high-impact operations.
• Investigated production incidents using logs, traces, reproducible test cases, and event histories, then delivered root-cause fixes, regression coverage, and concise operational runbooks for support teams.
Experience
Professional experience10+ years
Freelancer since2014
Languages
EnglishNative speaker