Recommendation Engine
Surface a related project or post automatically, not through manual linking.
Manually linking related content across 100+ projects and posts didn't scale and missed non-obvious cross-category connections.
A public prototype demonstrates the interaction; scope and data are intentionally limited.
Case Study
Before
Portfolio visitors needed personalized content suggestions across 100+ projects and blog posts. Manual related-content linking didn't scale and missed cross-category connections.
After
Automated content discovery across the portfolio. The engine surfaces non-obvious connections between projects and articles, increasing cross-page navigation and session duration.
How it's solved
Built a recommendation engine using collaborative filtering and content-based similarity scoring. The Python backend computes TF-IDF vectors for projects and blog posts, while the Next.js frontend displays contextual recommendations with confidence scores.
Trade-offs
Similarity scoring is TF-IDF based, not a trained embedding model — it can miss thematically similar content that shares little vocabulary.
Related Articles
Further reading on related topics from the blog
Building 100 Micro-Frontends: Lessons from a Solopreneur Journey
How I built and deployed 100 independently deployed micro-frontends on Vercel, the architecture decisions, trade-offs, and what I learned about scaling a one-person dev operation.
AI Integration Patterns for Web Developers: A Practical Guide
Practical patterns for integrating AI into web applications — from prompt engineering and streaming responses to building AI-powered tools that solve real user problems.
Essential AI Tools for Developers and Designers in 2025
A curated list of powerful AI tools that are transforming how developers and designers work, from OCR to design assistants.
Related Projects
More in Tech Domain · Book Dev