🧠 How I Think About Data

My MIT research on visual attention taught me that humans don't just see data—they construct meaning from patterns.

I solve data problems through a cognitive lens—across healthcare, federal consulting, and cloud platforms. These solutions aren't just technically sound, but designed with human cognition in mind: how attention functions, how memory organizes, and how understanding builds in layers.

This cognitive approach shapes everything I build: from dashboards optimized for natural eye movements to ML systems that explain their reasoning, and data products that work the way humans think.

Featured Projects

Fitness Dashboard

Self-Hosted Workout Intelligence

Attention patterns reveal signal in 14 years of messy data.

When my running data collided with daily dog walks, every workout got mislabeled as a “run”—even the 30 minute sniff walks. I built a full pipeline (auto csv export → λ → ) that classifies runs vs. walks using ML, with built-in model retraining for transparency.

Cognitive insight: Humans recognize patterns through contrast and repetition—this system leverages both to surface the real story in behavioral data.

Convoscope

Conversational AI Management

Multi-modal comparison designed for cognitive load management.

Hopping between AI providers while losing track of what worked where felt increasingly inefficient. Convoscope is my solution: one interface, multiple models (OpenAI, Anthropic, Google), persistent conversation history, and automatic topic extraction. It’s a flexible workspace for comparing outputs side-by-side.

Cognitive insight: Working memory can only hold 7±2 items—by offloading comparison to visual space, we free cognitive resources for actual thinking.

Beehive Tracker

Beehive Analytics Platform

Computer vision meets human memory to structure 4 years of visual documentation.

Four years of unlabeled bee photos became a living knowledge base by combining EXIF metadata, Google Cloud Vision, and weather APIs. The result: a knowledge graph with a query UI that surfaces patterns—when swarms happened, which seasons were productive, what weather preceded problems.

Cognitive insight: Human memory is associative, not chronological—this system mirrors how beekeepers actually recall and connect observations.

Cognitive Principles in Practice

👁️

Vision Drives Understanding

Design for scanning, not reading. Your brain decides what matters in 200ms.

See the Perception Guide →
🧠

Attention is Limited

Use it wisely. Humans can track 7±2 things—design within this constraint.

See Convoscope Example →
🔄

Patterns Beat Numbers

Humans think in stories. Show the narrative, not just the statistics.

See Fitness Story →

Resources & Guides

Built on Cognitive Foundations: These aren't just templates—they're tools designed with human cognition in mind. Each resource applies insights from visual perception, attention, and memory research to solve real data problems.

Executive Brief Template

Turn fuzzy threads into crisp, action-ready briefs with evidence links—designed for executive attention spans.

Dataset & Prompt Cards

Lightweight governance you can actually keep—structure, versions, and reproducible envs.

Browse all resources

🔬 Currently Exploring

AI Explainability Knowledge Graphs Human-AI Collaboration Attention Mechanisms in LLMs Data Product Design

Interested in discussing any of these? Let's connect