Law Zava
I engineer reliable infrastructure, reduce platform cost at scale, and lead technical teams to outship their peers.
Binary size, tail latency, memory predictability. When the runtime is the bottleneck, I reach for Rust, Zig, or C++ — not as a default, but when the numbers justify it.
ScyllaDB, Cassandra, and the operational reality of global state. Failure semantics, consistency trade-offs, and making distributed databases behave predictably under real load.
Small teams outperforming large ones. Clear intent, fast feedback loops, and async-first coordination over meetings and headcount.
Data residency, zero-trust architecture, and AI systems that satisfy regulators without crippling the product. Designed in from the start, not bolted on.
Reliability and cost discipline aren't at odds — they're the same engineering problem. Teams that understand their hardware, shrink their runtime dependencies, and make failure modes explicit end up with systems that are both cheaper and more reliable.
The best engineering organizations run on clear intent and fast feedback, not process overhead. I've seen five people with the right operating model outship fifty without one.
Structured red-teaming is a practical reliability discipline for distributed databases. Most catastrophic failures are compound scenarios nobody practiced, not black swans.
Local-first, hardware-aware architecture is becoming the default for high-reliability AI systems. The cloud-heavy pattern costs too much and fails too unpredictably for agentic workloads.
By early March 2026, the AI startup market looks less like a gold rush and more like a durable industry with clear pressure points. This post lays out where leverage sits, what buyers reward, and what durable execution looks like now.