3rd-year AI student at New Uzbekistan University. I build things that learn — from neural nets to production pipelines. Curious about everything, certain about very little.
I'm deep into machine learning, trying to understand how intelligence works — and how to make it useful in the real world.
class Yusuf: def __init__(self): self.name = "Yusuf Komilov" self.domain = "komilov.codes" self.year = 3 # at NUUz self.city = "Tashkent, UZ" self.loves = [ "machine learning", "glowing visuals", "late-night debugging", "strong tea", ] def current_focus(self): return { "learning": "transformers", "building": "RAG pipelines", "dreaming": "AGI someday", } yusuf = Yusuf() # open to collab ↑
University projects, experiments, and whatever I'm hacking on at the moment.
Fine-tuning a transformer for Uzbek NLP. Tackling low-resource language challenges with custom tokenization and domain-adapted pretraining.
An AI Q&A system for NUUz students. Indexes course materials, schedules, and docs using hybrid search + re-ranking for accurate retrieval.
Real-time artistic style transfer using CNNs. Optimized for CPU inference so it works without a GPU — built for fun, turned into a demo.
Interactive tool to explore transformer attention patterns across layers and heads. Helps students (and me) build intuition for how attention actually works.