Pinned Loading

1. Architecture_Analysis_Agent Architecture_Analysis_Agent Public

   Agentic AI system that analyzes software repositories, discovers architecture patterns, evaluates dependencies, identifies risks, and generates architecture review reports using planning, execution…

   Python

2. Researcher_Agent Researcher_Agent Public

   A research-first AI agent exploring agent architecture evolution using a shared-state Planner → Executor → Reflector → Replanner workflow.

   Python

3. self-attention self-attention Public

   A from-scratch PyTorch implementation of trainable self-attention, including Q/K/V projections, scaled dot-product attention, backpropagation, and contextual embeddings.

   Python

4. seq2seq-translator seq2seq-translator Public

   From-scratch English → Hindi neural machine translation using a vanilla LSTM encoder-decoder Seq2Seq architecture, implemented in PyTorch to study sequence-to-sequence learning, teacher forcing, an…

   Python

5. Word-Level-LSTM Word-Level-LSTM Public

   An LSTM language model built from first principles in NumPy for learning gated recurrent networks and long-term sequence modeling.

   Python

6. Word2Vec Word2Vec Public

   A clean, modular implementation of the Word2Vec Skip-Gram model built entirely from scratch in Python using NumPy. No high-level deep learning frameworks.

   Python