← HOME

/ BLOG

technical articles & systems deep-dives

How I built a Graph Neural Network recommendation engine from raw tensors and sparse matrices — and the systems/ML insights learned from LightGCN and 9 advanced optimizations.

#ai#ml#gnn#pytorch#recommendation#systems

Why a systems engineer spent the last few months building diffusion models from scratch — and what transferred, what didn't, and what's next.

#ai#ml#diffusion#pytorch#career#systems

A deep technical dive into building Carv, a programming language that compiles to C for ARM Cortex-M. Lexer internals, Pratt parsing, type inference, ownership tracking, C codegen, state-machine async/await, and the 11 bugs I haven't fixed yet.

#carv#compilers#embedded#systems#architecture

A deep technical breakdown of the Langflow unauthenticated RCE — how exec() with zero sandboxing allowed attackers to compromise AI pipelines within 20 hours of disclosure.

#security#rce#langflow#ai#python#code injection#sandboxing#supply chain

Zero is a systems language designed for AI agents. Structured JSON diagnostics, capability-based I/O, sub-10KB binaries. The idea sounds compelling until you look at how agents actually work with compilers today.

#zero#vercel#compilers#ai#systems#opinion#llm

A tactical guide for developers without formal experience: open source contributions, SaaS clones, freelance platforms, and building 'fake but real' projects that actually get you hired.

#career#portfolio#open-source#junior-developer#job-search