Now

Deep in the weeds of low-level systems programming. Building a custom Linux performance profiler in C++ to inspect program execution, stack sampling, syscall latencies, and cache/memory behavior beyond standard compiler abstractions.

Going deeper into modern generative AI infrastructure—moving past simple API wrappers to understand how agentic workflows, retrieval systems (RAG), and evaluation pipelines are designed and scaled in production.

Studying operating system internals, Linux kernel primitives, and distributed consensus mechanisms. Focusing on writing simple, resilient code that is easy to reason about and maintain.