📒 Gen AI for Developers — Video 1 Notes

1. Learned what GenAI for Devs means and how GPT (Generative Pretrained Transformer) works.

2. Understood Transformers architecture (embeddings, positional encoding, attention).

3. Learned what a Token is in AI and implemented basic tokenizer logic using tiktoken.js.

4. Explored Vector Embeddings and their role in semantic understanding.

5. Practiced Gemini API usage (@google/generative-ai) and compared it with OpenAI-style APIs.

   Basics Of GenAI + Tokens

   https://github.com/YashPandey1405/tokenizer-visualizer

📒 Gen AI for Developers — Video 2 Notes

1. Learned the importance of prompt engineering and how crafting effective prompts improves AI output.

2. Understood prompting styles including Alpaca, INST (LLaMA-2), FLAN-T5, and ChatML (OpenAI), with examples.

3. Explored system prompting techniques: zero-shot, few-shot, chain-of-thought (CoT), self-consistency, and persona-based prompting.

4. Practiced role-based prompting in ChatML and understood the importance of system, user, and assistant roles.

5. Implemented example prompts for translation, coding, and reasoning tasks to see differences between prompting styles.

   Master Prompting In AI

   https://github.com/YashPandey1405/system-prompts-genai

📒 Gen AI for Developers — Video 3 Notes

1. Learned the concept of Agentic AI, where AI can act autonomously using tools and step-based reasoning.
2. Implemented multiple practical agents/tools like weather, GitHub info, crypto prices, jokes, advice, activities, cat/dog facts, and exchange rates.
3. Practiced step-based AI workflow: START → THINK → TOOL → OBSERVE → OUTPUT for structured responses.
4. Built a NextJS + React frontend with dynamic tool mapping and JSON-based communication for scalable interactions.