Primary Technical Domains: AI & Deep Learning, CV, High-Throughput Web Platforms, Quantitative Modeling, Autonomous Agents.
🔬 Core Repository Clusters
1. 🧠 Frontier AI & Deep Learning from Scratch
C_Language_Deep_learning: Deep learning algorithms implemented from the ground up in ANSI C without external heavy runtime dependencies.
Numpy-Transformers: Complete Transformer architecture (Self-Attention, Multi-Head Attention, Positional Encoding, Feed-Forward) built strictly with raw NumPy.
Numpy-Cnn: Convolutional neural network training pipeline implemented from scratch in NumPy with forward/backward manual backprop.
Vision-transformers: In-depth Vision Transformer (ViT) implementation and architectural dissection.
Computer-vision-architectures: Comprehensive benchmark of modern computer vision backbones.
Thinking_Model: Architectural transformation pipeline enabling reasoning/thinking loops on standard LLMs.
package_code_antara & Antara_test: Modular implementation and continual learning benchmark suite for the ANTARA neural architecture under severe distribution shift (MoE, world models, cognitive governance).