Experiments and explorations
Not everything becomes a product. Some things are experiments - a way to test ideas, learn new approaches, and see what sticks.
LLM-Powered Developer Workflows
Exploring how large language models can assist with repetitive development tasks - code review, documentation generation, and architecture analysis. Not trying to replace thinking, just reduce friction.
Key takeaway
"Prompt engineering matters more than model size for specific tasks. Smaller, focused prompts consistently outperform broad instructions."
Android Architecture Patterns
Documenting and comparing different approaches to structuring Android applications - Clean Architecture, MVI, and simpler alternatives. Testing each pattern with real projects to understand trade-offs.
Status
Still in progress - notes and takeaways coming as this one develops.
AI-Assisted Code Review
Built a prototype tool that uses AI to provide initial code review feedback before human review. Focused on catching common patterns, not replacing judgment.
Key takeaway
"AI catches formatting and naming issues well, but architectural feedback still needs human context. Best used as a first pass, not a final verdict."