Problem
Smart X needed reliable multi-channel delivery at SaaS scale without locking every channel behind a single vendor bill. Provider APIs differ; new protocols show up when the business expands.
Featured case study
Multi-channel notification subsystem — Smart X
An in-house notification platform built for volume, cost control, and providers that keep changing.
Approach
Designed a fully flexible provider integration layer so new channels—including protocols like SMPP—plug in without rewriting the core pipeline. Used AI as a force multiplier to speed implementation while holding the line on quality through reviews, tests, and clear architecture.
Outcome
The subsystem processes 35M+ messages/day and cut provider costs by about 50% versus the prior path. Extensible integrations dropped new-provider work from roughly a week to about a day.
Selected work
More systems work
- Loyalty & segmentation (Smart X) — Led backend development for RFM and custom segments across 22M+ users using RabbitMQ, Redis, and background services in a distributed setup.
- API performance (Smart X) — Brought critical endpoints from about 5s down to roughly 300ms with query tuning, projections, and selective Dapper where it paid off.
- Realtime backends (Lucid Dream Games) — Services for multiplayer rooms supporting 20K+ concurrent players; internal voice-chat pipeline that replaced a third-party cost center; sustained 30K+ requests/min under 4GB RAM.
- Voice server (side systems project) — .NET voice service aimed at high concurrency with a small footprint and Docker-friendly deploy path.
Experience
Where the systems work happened
-
Designing and shipping backend modules for a SaaS platform: notifications at massive daily volume, loyalty/segmentation for tens of millions of users, query and API performance work, and migration toward more available microservice-shaped services. Heavy use of .NET, messaging, caching, DDD/vertical slices, and high test coverage.
-
Built and hardened realtime multiplayer backends—room concurrency, voice-chat pipeline economics, and latency/memory discipline under high request rates—while coordinating API design with the broader team.
-
Client networking and gameplay systems for titles with millions of players, tens of thousands DAU, and high concurrent sessions; Dockerized game-logic servers; mentored juniors and joined architecture calls. Earlier proof that scale and concurrency were already the job—not a later hobby.
Skills
Toolbox
Backend
.NET 8/9, ASP.NET Core, EF Core, Dapper, REST, Clean Architecture, Vertical Slice, DDD
Data
SQL Server, PostgreSQL, Redis, indexing, query plans, projections
Messaging & realtime
RabbitMQ, background workers, distributed events, high-concurrency services
Quality & delivery
xUnit, Moq, Bogus, Docker, Swagger/OpenAPI, code review
Education
Background
B.Sc. Computer Engineering — University of Science and Culture (Sep 2019 – Jul 2023). Cohort rank 1st · GPA 18.87 / 20.
English C1 (IELTS Academic overall 8.0).
Contact
Let’s talk systems
Open to backend and distributed-systems roles. Prefer email or LinkedIn.