Self-Directed AI Exploration — prompt engineering, iterative evaluation, and LLM app development (AI practice & learning).
Used Claude Code and OpenAI Codex for iterative prompt crafting and evaluation of AI-generated code outputs to improve correctness and relevance. Explored LLM behaviors for Java development workflows by comparing results across different AI coding assistants. Built a simple AI-powered dialogue application integrating Claude API, practicing multi-turn context handling and streaming responses. • Crafted structured prompts in Claude Code and iterated based on AI output evaluation. • Compared OpenAI Codex and Claude Code performance for code completion and translation. • Implemented a Spring Boot + Claude API dialogue app with multi-turn context and streaming. • Studied RLHF and AI safety concepts to inform how model outputs are assessed and improved.