CommerceSystemDemo
Production-style e-commerce backend with layered architecture, async PostgreSQL access, latency tuning, and complete metrics, traces, and logs for operational reliability.
vladislav.antonov — senior c++ developer · software architect
Platform Engineering · Backend Systems · Computer Vision
13+ years building high-performance C++ backends, computer vision pipelines, and distributed microservices. This front page highlights selected systems that show production architecture, reliability, and delivery discipline.
Production-style e-commerce backend with layered architecture, async PostgreSQL access, latency tuning, and complete metrics, traces, and logs for operational reliability.
Polyglot choreography-saga system (Go · Python · C++) that keeps a multi-service transaction consistent with no central orchestrator — transactional outbox, idempotent consumers, database-per-service, a synchronous gRPC leg over a Kafka backbone, and full metrics, logs, and traces.
Layered C++ ATM service with an HTTP adapter, token-based auth, and a red-to-green acceptance/BDD suite, demonstrating disciplined TDD and clean architecture in a realistic banking workflow.
C++17 mobile-operator billing engine built from an explicit specification, with domain modeling, warnings-as-errors builds, and a full ctest quality gate covering the tariff rules.
Offline C++20 VoIP analyzer that reconstructs SIP dialogs (RFC 3261) and correlated RTP streams (RFC 3550) from PCAP, emitting CDR-style JSON with running jitter, packet loss, and E-model MOS estimates.
Algorithm-focused CV project with CUDA, OpenCL, Vulkan, and OpenGL compute implementations, demonstrating optimization across multiple GPU backends.
C++ implementation of the CVPR 2011 normalized-sparsity blind-deconvolution algorithm, modernized to C++20 as a reusable static library with bit-exact parity to the original port and a documented QA report.
C++ CI reference with static analysis, sanitizers, valgrind, and automated quality gates, covering the engineering discipline needed to keep large systems maintainable.