About me

AI-based trading platform with full ownership of backend architecture, data infrastructure, and transaction processing.
Rust, Go, Node.js, ClickHouse, PostgreSQL, ScyllaDB, Redis, GCP. Led the architecture of a time-series data platform spanning real-time ingestion, analytical processing (ClickHouse), and cold storage (GCS data lake). Architected a transaction processing system with a custodial ledger (idempotency, reconciliation, replay). Improved critical-path efficiency (-20% CPU, -15% memory).

Neo-banking platform (Revolut-like) with focus on platform architecture, auth/KYC, and distributed integrations.
Python, FastAPI, Keycloak, PostgreSQL, RabbitMQ, AWS. Drove architectural decisions for API gateway, auth/KYC services, and event-driven infrastructure. Designed an internal event-bus framework (Layerbus). Built KYC pipelines integrated with KYCAID and Sumsub.

Backend and data systems for ML-heavy client projects. Designed ETL pipelines for video processing supporting ML training/inference, reproducible training/debugging pipelines with dataset versioning, and speech-processing pipelines (TTS, voice conversion).

Owned backend systems for ordering and payments platform (Chowtime, Bookme). Built end-to-end QR-based ordering with POS integration; designed background processing and synchronization flows; integrated Zoom, Twilio, Google Calendar.

Owned backend architecture and delivery; maintained system performance and data integrity in production.

Payment systems aggregator and terminal infrastructure. Designed integrations with 40+ payment providers and banks; built an offline transaction processing pipeline connected to a central ledger; implemented reconciliation and fault-tolerant processing; designed crypto and fiat payment flows.

Built Pub/Sub-based service communication. Implemented async processing and backend optimizations.

Designed and delivered backend systems across healthcare, e-commerce, and trading. Built APIs, data models, integrations, and async processing pipelines.