hyper07 - Overview

Hi there 👋 I'm Kibaek.

🌍 Senior Full-Stack Software engineer | ML & AI enthusiast

Welcome to my GitHub! I build tools, experiments, and frameworks in machine learning, generative AI, and modern backend systems. I enjoy learning in public and sharing open-source projects that help others build and explore.


About Me

  • I’m currently working on Generative AI & ML projects, including a full-stack GenAI Playground with PyTorch, FastAPI, and Docker.
  • I focus on designing scalable, low-cost LLM system architectures, emphasizing efficient inference, modular pipelines, and production-ready deployment.
  • I enjoy building end-to-end machine learning workflows — from data preprocessing to model training, API design, evaluation, and deployment.
  • Ask me about Python, ML/AI systems, LLM infrastructure, backend APIs, and project architecture.

Featured Projects

Here are some of my key open-source repositories:

🌐 Full-stack & AI

  • gen-ai – A full-stack Generative AI Playground with modern deep-learning models, a FastAPI backend, Dockerized setup, and Jupyter support. (Explore and experiment with LLMs and vision models)
  • Deep-Learning-LLM-Wounded-Treatment – A practical deep learning project combining CNNs and LLM workflows for domain-specific medical applications.
  • AML Detection Platform (NDA) – Built a Dockerized, end-to-end AML system using unsupervised anomaly detection, PCA, and ensemble risk scoring with analyst-facing dashboards.
  • Exo Proxy Fortress – Designed and built a secure, high-performance LLM load balancer and proxy layer for distributed AI inference across Exo clusters and Ollama servers.
    Implements intelligent request routing, model-aware load balancing, API key–based security, rate limiting, and application-level cluster management without Kubernetes.
    Extended the system with RAG (Retrieval-Augmented Generation), vector search, and Redis-backed KV caching to significantly reduce latency and inference cost.
    Fully Dockerized with FastAPI, Gunicorn, Redis, MongoDB, ChromaDB, and Nginx for production-ready deployment.

🛠️ Tech Stack

Languages: Python, JavaScript/TypeScript, PHP, SQL
Frameworks: PyTorch, FastAPI, Flask, React, Symfony
Tools & Platforms: Docker, Jupyter, GitHub Actions, REST APIs
AI & ML: Transformers, CNNs, GANs, Diffusion Models, XGBoost, HBOS, Isolation Forest, ECOD


🌱 What I’m Learning

  • Large-scale LLM training & fine-tuning
  • Cost-efficient LLM inference and system optimization
  • Production-grade ML and AI deployment best practices

Languages and Tools:

aws canvasjs chartjs css3 docker express firebase flask git html5 javascript jenkins jest kubernetes linux mariadb mongodb mssql mysql nextjs nginx nodejs pandas php postgresql postman python pytorch react reactnative redis scikit_learn seaborn swift symfony tensorflow