About

I got good at engineering by breaking things and staying long enough to fix them properly. Eight-plus years of backends later, I still think the interesting work starts after the demo.

What I do

I build services with clear contracts and boring, trustworthy telemetry. Right now that's email deliverability and sequences at Apollo.io. Before that, messaging and notifications at Snapdeal: counts that had to match, cost that had to come down, bots that had to stop being a mess.

I write about the same world. Retries that stampede. Metrics you can't trust. The human side of a page. Questions that make you less junior. Why any of this is worth the cost.

How I think about the work

Seniority is not a bigger pattern catalog. It's seeing what happens twice, late, partially, or during recovery, before it hits someone at 3 a.m.

It's also how you show up. Calm when the room wants a hero. Risk said early. Docs that still make sense next year. Time spent with people earlier on the path.

Focus

• APIs and services: contracts, performance, migrations
• Reliability and observability: SLOs, metrics, logs, traces
• Data systems: PostgreSQL, MongoDB, Elasticsearch, Kafka

Stack

Ruby, Go, Java, Python. PostgreSQL, MongoDB, Redis, Aerospike, Kafka, Kubernetes. Prometheus, Grafana, Datadog, OpenTelemetry.

Principles I actually use

• Design for on-call: idempotency, backpressure, timeouts
• If you can't count it cleanly, you can't operate it
• Small reversible steps beat clever rewrites
• Leave the next human better notes than you got

timeline → · work → · now → · @theybanjan · writing →