Focus 01
Bias detection
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Measuring disparity across protected attributes before deployment.
Currently
Placeholder lede. I build tools that make machine learning systems legible — auditing datasets before they become models, and models before they become decisions.
Focus 02
Interpretability
Sed do eiusmod tempor incididunt ut labore. Attribution methods that non-technical stakeholders can actually read.
Focus 03
Monitoring
Ut enim ad minim veniam, quis nostrud exercitation. Tracking fairness metrics as they drift in production.
Placeholder text. The Projects tab is the main event — start there.
Selected work
Projects
- Responsible AI
- Data viz & communication
- Earlier work
Responsible AI
Bias detection in hiring models
Surfacing demographic disparities in candidate scoring
Placeholder copy. Lorem ipsum dolor sit amet, consectetur adipiscing elit. A framework for surfacing demographic disparities in candidate scoring before a model ever reaches a recruiter, with per-group threshold analysis and a plain-language report for hiring teams.
View on GitHub
Responsible AI
Explainability toolkit
Attribution methods built for non-technical stakeholders
Placeholder copy. Sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Feature attribution for tabular and NLP models, wrapped in an output format that a policy team can read without a stats background.
View projectResponsible AI
Dataset auditing pipeline
Representation checks that run before training starts
Placeholder copy. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip. Automated checks for representation gaps, label noise, and proxy variables, run as a pre-training gate in CI.
Data viz & communication
Model monitoring dashboard
Watching fairness metrics drift in production
Placeholder copy. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore. A lightweight dashboard that tracks group-wise error rates over time and raises a flag when the gap widens past a set tolerance.
View project
Data viz & communication
Interpretable NLP embeddings
Making semantic clustering visible
Placeholder copy. Excepteur sint occaecat cupidatat non proident. An interactive projection of learned representations that surfaces unexpected clustering along protected attributes.
Earlier work
Sentiment classifier
An early look at class imbalance in social media data
Placeholder copy. Sunt in culpa qui officia deserunt mollit anim id est laborum. An undergraduate project on class imbalance in social media sentiment datasets — kept here because it is where the interest started.
Background
About
Placeholder paragraph. Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris.
Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
MethodsFairness metrics · counterfactual testing · SHAP / LIME · dataset auditing
ToolsPython · PyTorch · scikit-learn · pandas · D3
WritingPlaceholder — talks, papers, and posts go here
