Projects
Systems That Think Like Humans Think
Each project applies cognitive science principles to real-world data challenges. From visual attention patterns to memory organization, these aren't just technical builds—they're explorations of how human cognition can shape better systems.
🎯 Currently Shipping
Active projects where cognitive principles meet production code
Fitness Intelligence Platform Behavioral Patterns
Cognitive Principle: Pattern recognition through contrast and repetition
When 14 years of running data collided with dog walks, I discovered that human behavior naturally clusters into 3 categories—not the 10+ my app offered. The ML classifier doesn’t just sort workouts; it mirrors how humans actually categorize effort levels when cognitive load is high (like during exercise).
Technical: Python, AWS Lambda, MySQL, Streamlit
Cognitive Insight: The 7±2 rule emerges naturally in behavioral data
→ View Project | → Live Demo | → GitHub | → Documentation
Convoscope: Multi-Modal AI Comparison Model Evaluation
Cognitive Principle: Cognitive load management through progressive disclosure
Comparing AI outputs revealed a fundamental limit: humans can effectively compare 3 options, tolerate 5, and get overwhelmed by 7+. This interface respects working memory constraints while managing infinite AI possibilities.
Technical: Streamlit, OpenAI/Anthropic/Google APIs, PostgreSQL
Cognitive Insight: Side-by-side comparison reduces cognitive load vs. sequential evaluation
→ View Project | → Live Demo | → GitHub | → Documentation
Beehive Knowledge Graph Metadata Integration
Cognitive Principle: Human memory is associative, not chronological
Four years of photos became queryable knowledge by modeling how beekeepers actually recall events—through associations (weather→outcomes) rather than dates. The system’s 7 relationship types emerged naturally from how humans connect observations.
Technical: Neo4j, Google Cloud Vision, Python, Weather APIs
Cognitive Insight: Domain expertise is about relationships, not isolated facts
→ View Project | → Live Demo | → GitHub | → Documentation
🔬 Research & Exploration
Where cognitive science theory meets data science practice
Academic Citation Network Prediction Knowledge Graph
Cognitive Principle: Attention mechanisms in knowledge discovery
Using TransE to predict missing citations in a network of 8,000 papers. The model learns “attention patterns” in research—which ideas researchers notice and which they overlook, despite relevance.
Technical: Graph Neural Networks, PyTorch, Neo4j, Semantic Scholar API
Cognitive Insight: Academic attention follows predictable patterns, just like visual attention
→ Data Story | → Live Demo | → GitHub | → Documentation - Coming Soon | → Jupyter Notebooks - Coming Soon
đź“š Foundational Builds
Earlier projects that established my cognitive approach
IoT Temperature Sensor Fleet (2021)
Cognitive Principle: Real-time data needs real-time comprehension
Built a distributed sensor network that taught me: streaming data is only valuable if humans can process it at the speed it arrives. Led to my focus on progressive disclosure and attention management.
Digital Portfolio Workshop Series (2024)
Cognitive Principle: Information architecture as cognitive architecture
Teaching others to build portfolio sites revealed how cognitive principles apply to personal branding and content organization.
→ Part 1 of 4 Series → Part 2 of 4 Series → Part 3 of 4 Series → Part 4 of 4 Series
The Cognitive Thread
Every project here demonstrates the same truth: the best technical solution considers human cognition first.
Whether it’s chunking workout data into 3 categories, limiting AI comparisons to working memory capacity, or organizing beehive observations like human memory works—the pattern is consistent. Technology succeeds when it aligns with how humans naturally think.
Want to discuss building something that works the way humans think? Let’s connect →