Ground Zero
New phase, new experiments.
Look, I genuinely believe the best way to push AI forward is by sharing everything — the wins, the failures, the weird stuff in between. Not because I’m some altruistic genius, but because that’s how we all learn faster. You build on my experiments, I build on yours, someone else combines both and discovers something wild. That’s the cycle.
So yeah, expect frequent drops of whatever I’m tinkering with — code, thought dumps, deep dives into mechanisms that caught my attention. If it helps even one person skip a week of debugging or sparks a new idea, mission accomplished.
The Journey (aka How I Got Here)
Started out obsessed with cybersecurity post-bachelors. Like really obsessed — pentesting, bug bounties, network exploitation, the whole nine yards. Somewhere along the way during masters, I got sucked into the AI rabbit hole and never climbed back out.
Spent a ton of time experimenting with:
Building multi-agent pipelines
Orchestrating AI workflows that actually worked
Deep diving into LoRA adaptations (which ones work, which ones are overhyped)
Combining models for depth estimation + object detection
Basically anything that seemed interesting at 2am
Wrote a bunch on Medium during this phase — some technical breakdowns, some philosophical tangents about life design and tech, some straight-up AI news commentary. The reach was decent, but here’s the problem: paywalls. I don’t want my experiments locked behind subscriptions. Knowledge should flow freely, especially when we’re all just figuring this stuff out together.
What I’m Up To Now
Currently at Mako making kernels faster. Building agentic pipelines and fine-tuning LLMs with RL to generate optimized GPU code across CUDA, HIP, Triton, and Tenstorrent. Automating what used to take specialists weeks.
After bouncing from cybersecurity → AI/ML → RL, I’m at this point where I want to share experiments that actually matter. Stuff inspired by nature, human behavior, intersections nobody’s really explored yet.
What to Expect Here
This Substack is my new playground for:
RL experiments (slightly obsessed, not gonna lie)
Messing with LLMs in unconventional ways
Vision model tinkering
Whatever strikes my curiosity that week
Probably some 3am thought dumps that may or may not make sense
Just open experiments and honest documentation of what works (and what spectacularly doesn’t).
My Past Work
All previous articles are still up on Medium if you want to dig through the archives: 👉 medium.com/@teendifferent
Fair warning: it’s a mix of technical deep dives, AI, RL cybersecurity nostalgia, and random philosophical detours. Navigate accordingly.
Let’s Build Together
That’s it for the intro. If you’re into running experiments, breaking stuff to understand it better, and sharing what you learn — you’re in the right place.
Stay tuned. Let’s explore this together.