Neural network exercises and a transformers textbook for gen_ai_engineer
By RepoJournal , from @anurag629's public GitHub activity
Backfilled
1 person shipped this
Anurag Verma built out a structured learning path with hands-on exercises, an autograd engine, and a comprehensive 10-chapter transformers book across five commits to gen_ai_engineer.
The core work was a set of neural network exercises [1] paired with a simple autograd implementation in micrograd.py to support gradient calculations. The exercises themselves cover six problems: verifying gradients, solving XOR, plotting training loss, comparing learning rates, classifying the moon dataset, and comparing results with PyTorch. Alongside this, Anurag added a .gitignore [2] for Python, IDE, and model files to keep the repository clean.
The learning path expanded with interview preparation terms and concepts added to Day 1 [3], followed by a full Day 2 module [4] on language modeling and NLP covering bigram models, neural bigrams, and MLP language models. The final piece was a substantial Day 3 book [5] on transformers spanning 10 chapters and over 2300 lines, paired with visualizations across the curriculum: eight concept diagrams for Day 1 (activation functions, computational graphs, gradient descent, loss landscapes) and generated plots for Day 2 showing bigram counts, training loss, and embeddings.
One email a day. Unsubscribe in one click.
A short briefing every day anurag629 ships something — in about 3 minutes.
One email a day. Unsubscribe in one click. Read a past issue →
References
- [1] Add exercises and micrograd implementation for neural networks anurag629/gen_ai_engineer
- [2] Add .gitignore for Python, IDE, and model files anurag629/gen_ai_engineer
- [3] Add interview prep terms and concepts to Day 1 anurag629/gen_ai_engineer
- [4] Add Day 2: Language Modeling & NLP - bigram, neural bigram, MLP language model anurag629/gen_ai_engineer
- [5] Add visualizations for Day 1 & Day 3, create Day 3 Transformers book anurag629/gen_ai_engineer
