infra engineering is underrated
real-time systems and distributed consensus are underrated career paths right now — the AI talent drain is creating real opportunity in infrastructure engineering.
high-frequency data processing, streaming analytics, and sub-millisecond latency optimization involve constraints that are physical and measurable — you can't prompt-engineer your way past hardware limits.
performance engineering at scale is genuinely hard — consensus algorithms, distributed state machines, and fault tolerance push the boundaries of what's computationally possible.
these fields are also more resilient to AI automation because the problems are tightly coupled to hardware, network physics, and system-level tradeoffs.
if you're looking for deep technical work outside the ml hype cycle, infra engineering has never had more opportunity and less competition for talent.