Event-driven payment infrastructure on Java and Kafka, holding sub-50ms p99 latency under sustained load.
We do not publish client names
The platform needed payment and order APIs that stayed correct under heavy concurrent writes, integrated third-party payment providers, and handled card data to PCI-aligned standards, without latency degrading as volume grew.
Event-driven processing on Kafka: producers stream payment and order events with partitioning for parallel consumption, and consumers run a matching engine with price-time priority. Idempotency keys on every write. PostgreSQL with partitioning, composite indexes and optimistic locking for ACID correctness under concurrent payment writes. Redis caching using Sorted Sets and Lua scripts. OAuth2 and JWT throughout, with third-party provider integrations via API Gateway and Lambda.
The system sustains 10,000 transactions per second at under 50ms p99. Redis caching cut PostgreSQL read load by 85%. Six months in production with zero critical bugs. Terraform, Jenkins and GitHub Actions drive automated builds, container-based integration tests and static analysis quality gates.
Our own product, designed and built end to end: QR check-in with no hardware, GST-compliant invoicing, WhatsApp reminders and live video classes inside the operator's own app.
Four microservices powering retrieval, generation, configuration and memory as a knowledge agent inside a larger agentic system.
Distributed Go services processing regulatory rules for search, sharing and management on one platform.
Describe the system and the constraint you have hit. You will get a technical reply, not a sales sequence.
