Projects
If I can't measure that it works, I don't ship it. The rule holds for everything here: engineering systems, peer-reviewed research and practice labs. Filter by type, or browse them all, newest first.
-
7PH Graph: A Knowledge Graph of a Competitive Metagame
A live knowledge graph of a competitive Magic format: 107 events, 4,591 decks, every chart backed by provenance and statistical guards.
PythonCypherGradioPlotly
-
Digital Twin
A conversational agent answers questions about my work, grounded in a curated knowledge base. A second model reviews every answer before a visitor sees it.
PythonChromaDBLiteLLMGradio
-
AI-JIE: Extraction and Evaluation Pipeline
A chain-of-thought extraction pipeline turned 3,892 job postings into validated, intent-classified records. Human review scored the final prompt at 4.11 of 5.
PythonOpenAI APIPydanticasyncio
-
LLM Engineering Lab
11 projects spanning retrieval, fine-tuning and autonomous agents. The flagship ensemble predicts product prices with a mean absolute error of $29.95.
PythonQLoRAChromaDBModal
-
Job Intelligence Engine
An extraction pipeline turns 6,100 job postings into ranked role recommendations. A judge model scores the extraction layer's accuracy.
Pythonscikit-learnSBERTXGBoostStreamlit
-
MLB Analytics with SQL
A reusable SQL workflow answered 4 questions on 150 years of baseball data, from college talent pipelines to payroll overperformance.
PostgreSQLSQLPython
-
Python Labs
3 lab collections, one progression: object-oriented Python systems, exploratory data analysis and classical machine learning, the practice ground behind the shipped systems.
Pythonpandasscikit-learn
-
Seasonal altitudinal migration in rainforest birds
16 years of counts showed rainforest birds shifting uphill each summer and back each winter, the first system-wide measure of partial altitudinal migration.
RJAGSN-mixture models
-
Forest gap effects on tropical birds
Forest gaps reshaped bird assemblages across 1,148 captures of 81 species, while total abundance stayed level between gaps and closed canopy.
RGLMs
-
Physiological stress and rainforest mammal declines
A model chain from microclimate to physiology to demography showed heat stress and foraging limits driving possum declines, species by species.
RJAGSBiophysical models
-
Climate, foliage chemistry and herbivory
Hierarchical models traced climate and geology through soil and leaf chemistry to herbivory. Single soil nutrients predicted little once geology was controlled.
RJAGSHierarchical models
-
Climatic drivers of rainforest bird change
Warming and shifting rainfall drove opposite trends in lowland and upland birds across 47 species. Cyclones and droughts had only marginal effects.
RJAGSRemote sensing
-
Ringtail possum viability forecast
A hierarchical Bayesian model on 30 years of surveys forecast possum collapse by 2050. It fed a national protection nomination.
RJAGSForecasting
-
Community reshuffling under elevational shifts
Simulated uphill shifts of 7,613 assemblages predicted mass local extinctions of upland species, mapping the escalator to extinction along the elevational gradient.
RSpatial forecasting
-
Rainforest bird declines
17 years of bird monitoring showed upland populations nearly halving. The evidence supported protection nominations for 14 rainforest species.
RGLMsShiny
-
Predicting abundance from environmental suitability
An ensemble of 9 algorithms predicted local abundance from environmental suitability, explaining 55% of deviance on average across 50 rainforest species.
REnsemble MLMaxEnt