AI Developer

Notes on building AI-powered products through code. This covers working directly with AI APIs, model selection, prompt engineering, and production patterns.

What's Inside

LLM Stack — The model lineup I use, fallback chains, streaming, and system prompt design.

Google AI Studio — Gemini model families, free tier limits, API functions, and code examples for every capability.

RAG Pipeline — Retrieval-Augmented Generation architecture for grounding LLMs in your own data.

My AI Stack

Layer	Tool
Primary model	Gemini 2.5 Flash / 3.x Flash
Fallback chain	gemini-2.5-flash-lite → gemini-3.1-flash-lite
SDK	`@google/generative-ai`
Embeddings	`gemini-embedding-001`
Vector store	Pinecone / pgvector
Streaming	Server-Sent Events (SSE)
Framework	Next.js App Router + Route Handlers

Quick Start

import { GoogleGenerativeAI } from "@google/generative-ai";

const genAI = new GoogleGenerativeAI(process.env.GEMINI_API_KEY!);
const model = genAI.getGenerativeModel({ model: "gemini-2.5-flash" });

const result = await model.generateContent("Explain RAG in 2 sentences.");
console.log(result.response.text());

See also: AI Automation for low-code n8n workflows.