The snake animation is generated automatically by GitHub Actions. The first run may take a little while to publish.
Hey! I'm Sepehr — an AI enthusiast who learns by building.
I'm curious about what's happening under the hood: how models learn, how agents use tools, and how to turn an idea into a real, working project. I especially enjoy experimenting with local AI, so models can run on my own machine instead of relying on a hosted API.
- 🧠 Exploring AI, machine learning, and neural networks
- 🛠️ Building Nova, a local-first AI agent powered by Ollama
- ♟️ Experimenting with chess engines and AI decision-making
- 🔬 Learning by creating practical ML experiments
- 💻 Exploring C/C++ and how software works closer to the machine
- 📐 Interested in advanced mathematics and the ideas behind machine learning
- 🌍 Improving my English alongside programming and AI
Curiosity is the roadmap. Building is how I learn.
| Learning how neural networks represent patterns, how training works, and how to build experiments from the ground up. | Exploring the mathematical ideas that power AI: algebra, functions, probability, linear algebra, and eventually calculus. |
| Understanding memory, performance, and how programs work at a lower level. | Building stronger English skills to read documentation, understand research, and communicate technical ideas. |
A local-first AI agent built around Ollama. Nova can plan tasks, use registered tools, inspect tool results, re-plan after failures, and work with user-controlled local memory.
|
♟️ Chess Model
|
A machine-learning practice project using histopathology image data to explore dataset handling, model training, and prediction workflows. For learning and experimentation, not clinical use.
|
I keep experimenting with AI, interfaces, automation, and ideas I want to understand better.
|
