Backend engineer building distributed systems for B2B SaaS.
I work on backend problems where throughput, correctness, and product constraints collide: spike traffic, multi-tenant fairness, streaming pipelines, and reliability under failure.
Over the last 5+ years at Channel.io, I have worked on CRM and marketing systems that process 100M+ messages per month. I care about systems that hold up in production, and I write about the trade-offs behind them here.
What you will find here
- Real production problems, not toy examples
- System design decisions and their trade-offs
- Notes on distributed systems, databases, and backend architecture
- Writing that reflects how I think and debug as an engineer
A few things I have worked on
- Built a bulk processing system that handled 1k/s spike traffic with better fairness and failure isolation
- Replaced batch-style aggregation with a DynamoDB CDC pipeline and reduced database load from 297 TPS to 36 TPS
- Introduced ClickHouse for more expressive CRM targeting queries
- Investigated and fixed production issues such as monolith OOMs and DynamoDB GSI backpressure propagation
- Designed shared reliability patterns including idempotency, retries, and rate limiting
Start here
- Handling Spike Traffic Reliably: Improvement
- More Precise Customer Targeting with ClickHouse
- Handling Spike Traffic Reliably
- Database Internals
- Transactions and Consensus in Distributed Systems