Slymi. Building the Future of AI.

Driven developer crafting efficient, innovative AI solutions, including the next-generation Apertis LLM.

Explore Apertis

About Me

I'm Slymi, a developer deeply immersed in the world of Artificial Intelligence and Large Language Models. My primary focus is pushing the boundaries of efficiency and capability in AI, exemplified by my work on the Apertis LLM architecture. I thrive on complex challenges and transforming ambitious ideas into functional, high-performance code.

Beyond AI, I possess strong skills in Python development and modern web technologies (HTML, CSS, JavaScript). I'm constantly learning and driven by the potential to create technology that makes a significant impact.

"The worst thing you could possibly do is give up on trying everything that you are able to do. Throwing away the potential because you tell yourself you can't."

Featured Project: Apertis LLM

A novel, efficient, and simple Multimodal LLM Architecture.

Apertis is designed from the ground up to overcome limitations in existing LLMs, focusing on linear-time processing, memory efficiency, native multimodal support, and ease of use. It integrates cutting-edge concepts into a cohesive and powerful architecture.

View on GitHub

Core Architectural Innovations:

⚙️

Selective Linear Attention (SLA)

Achieves O(n) time complexity for efficient processing of long sequences, adapting attention dynamically.

🧠

Adaptive Expert System (AES)

Inspired by MoE, it dynamically routes tokens to specialized experts for efficient parameter usage and computation.

💾

State Tracking Recurrent Cell (STRC)

Combines RNN-like state tracking with transformer capabilities for efficient inference and context management.

🖼️

Unified Multimodal Encoder (UME)

Seamlessly processes and integrates both text and image data within a joint embedding space.

Professional Training & Interface

Apertis includes a user-friendly, Google AI Studio-inspired UI built with React/FastAPI, featuring YOLO-style real-time training visualization and cross-platform compatibility (Windows, macOS, Linux via Docker).

Other Projects

AI Macro (Bee Swarm Sim)

An innovative AI-powered macro demonstrating image processing and automation techniques for game interaction.

Python AI Image Processing

Calorie AI (Upcoming)

A planned project focused on applying AI for personalized calorie tracking and nutritional insights.

Python AI Data Analysis

Web Development

Various web projects demonstrating proficiency in front-end technologies and creating engaging user interfaces.

HTML CSS JavaScript

Technical Skills

AI & Machine Learning

  • Python (Advanced)
  • LLM Architecture Design (Apertis)
  • PyTorch Basics
  • Multimodal AI Concepts
  • Image Processing
  • AI Project Leadership

Web Technologies

  • HTML5
  • CSS3 (Flexbox, Grid)
  • JavaScript (ES6+)
  • React Basics (from Apertis UI)
  • FastAPI Basics (from Apertis UI)

Tools & Platforms

  • Git & GitHub
  • Docker
  • Visual Studio Code
  • Linux/Unix Basics
  • Windows

Get In Touch

Interested in collaborating or have questions about Apertis or my other work? I'd love to hear from you.

[email protected]