|
|
|
“Deep learning is just applied calculus, but please don't tell the prompt engineers.”
Sometime around learning what a gradient actually was, my brain fundamentally broke, and I permanently lost the ability to accept "it just works" as a valid engineering answer.
While normal developers are out building aesthetically pleasing web apps, maintaining a healthy work-life balance, and enjoying their weekends, I am usually in a dark room manually implementing automatic differentiation engines and arguing with memory allocators. I firmly believe that high-level APIs have made the industry soft. True enlightenment isn't found in a pip install command—it's found somewhere between a core dump, a perfectly formulated Markov Decision Process, and a three-day debugging bender.
C++We have a highly toxic, codependent relationship, but manual memory management builds character.CLike C++, but with absolutely zero safety nets. I write in C when I truly want to feel alive (or completely miserable).Mathematical OptimizationThe only true framework. Everything else is just syntax applied to matrix calculus.
PythonMy safety blanket. I begrudgingly use this when the C++ compiler has successfully broken my spirit for the day.JavaI just tried it. WritingAbstractSingletonProxyFactoryBeanmakes me feel like I'm trapped in a corporate simulation, but technically, it runs.Docker & KubernetesI put my chaotic low-level code inside little isolated boxes so it can't escape and hurt the rest of the file system.Terraform / PrometheusBecause sometimes you have to prove you can orchestrate systems, not just write math equations on a whiteboard.
JavaScriptI know it. I hate it. We do not speak of it here.
🤝 You are a Recruiter / Hiring Manager
Hi! I am a highly motivated engineer passionate about robust ML systems, hardware-aware AI, and scalable backend architecture. I promise I write clean, documented code when I'm on the clock. Check my pinned repositories for actual, production-grade logic.
🧮 You are a fellow Low-Level/Math Masochist
Welcome, brother. If you also enjoy writing custom tensor libraries from scratch because PyTorch "abstracts too much," we need to be friends. Let's talk about dynamic autodiff engines and convex analysis.
🎨 You are a Frontend / Web Developer
I am terrified of you. Please keep your CSS, your 40,000 node_modules, and youruseStatehooks away from me. I will write an entire operating system before I try to vertically center a<div>.
| Metric | Current Status |
|---|---|
| 🐛 Bugs Written | Overflow Error |
| 🧠 Tensors Reshaped | [1, 512, 512, 3] -> [1, -1] -> Panic |
| 💥 Segfaults Survived | Countless |
| ☕ Blood-Caffeine Level | Critical |
| 🔋 Sanity Remaining | [██░░░░░░░░] 20% |
| 🩸 C++ Stockholm Syndrome | [█████████░] 99% - I CAN FIX IT |
| 📉 Gradient Descent | [████████░░] 80% - STUCK IN LOCAL MINIMA |
