· 1 min read
My 2025 Wrapped: What I Built, Shipped, and Learned in AI
A look back at the AI side projects I built in 2025 and the technologies that shaped how I think about shipping AI into production.
By Tanner Goins
This year, I spent a lot of time building AI projects outside of work and going deeper into the tools that shape modern AI systems.
AI Side Projects I Built in 2025
- Neural networks trained on real user engagement data to learn behavioral patterns and predict what content people would enjoy
- Growth-focused personalized content creation tool combining deterministic rules with ML to recommend what to create and when
- AI stock trading assistant designed for decision support, notifications, and strict guardrails
- Unity 3D AI simulations, exploring how agents move between tasks and systems in a more spatial, visual way
- Agentic personal finance application with automated categorization, forecasting, NLP, and LLM-driven workflows
AI Technologies I Focused On
- LangChain and LangGraph for structured multi-step LLM orchestration
- Agentic patterns, stateful reasoning, and event-driven AI workflows, including exploring the tradeoffs between single prompt and step-based LLM designs
- Brain.js in TypeScript for custom feed-forward networks, classifiers, and time-series predictors
- AI observability and auditing model decisions
- Defining strict boundaries between deterministic logic and LLM interpretation
Looking Ahead
Each project pushed my understanding of how AI can deliver real business value, whether through better predictions, faster insights, or smarter workflows. Looking forward to pushing that further in 2026.