I study how information behaves in code: in commits, in adoption curves, in systems that look chaotic until you measure them properly. Most of what's below started as a clean hypothesis. Some of it ended with the data winning instead of me. I think that's the more interesting outcome to publish.

difference-in-differences quasi-experimental git mining 638k commits

Does GitHub Copilot change how people commit, causally rather than just correlationally?

flowchart LR
    A[GH Archive<br/>10 repos, 2018-2024] --> B[Commit extraction<br/>638,000+ commits]
    B --> C[Feature engineering<br/>files / churn / inserts]
    C --> D[Panel: repo x week<br/>fixed effects]
    D --> E[Pre-trend F-test]
    D --> F[DiD + event study]
    D --> G[Commit-message DiD<br/>mechanism check]
    E --> H[Result]
    F --> H
    G --> H

Effect concentrates in existing contributors, not newcomers: people write smaller, more atomic commits once an assistant is doing the typing. A commit-message DiD adds the mechanism check, with fix-commit share falling and refactor share rising post-adoption. HC3 robust SEs throughout; IEEEtran paper drafted, targeting MSR 2027.

staggered-adoption DiD sun-abraham estimator bigquery 89 repos

Does adding a CODEOWNERS file causally change how fast pull requests get reviewed?

flowchart LR
    A[GH Archive<br/>~21TB via BigQuery] --> B[89-repo panel<br/>staggered adoption]
    B --> C[Sun-Abraham<br/>event-study DiD]
    C --> D{Result by<br/>horizon}
    D -->|0-18mo| E[Null on<br/>PR closing time]
    D -->|23-24mo| F[Signal, flagged<br/>as confounded]
    B --> G[CODEOWNERS coverage<br/>parsed at treatment date]
    G --> H[Bimodal split:<br/>14 repos ≤10% / 8 repos ≥90%]

shannon entropy leave-one-repo-out CV honest null result

Does Shannon entropy in commit histories predict upcoming software releases?

flowchart LR
    A[Commit histories<br/>9 repos] --> B[Shannon entropy<br/>per commit window]
    B --> C[Naive train/test split]
    C --> D[72% accuracy]
    B --> E[Leave-one-repo-out CV]
    E --> F[AUC 0.47, chance level]
    B --> G[Confound check]
    G --> H[Commit volume vs entropy<br/>Spearman r = 0.817]

Status: Published as a negative result with a documented confound, not a quiet repo nobody talks about. The methodology is the part worth reading.

fairness-aware ml demographic parity equal opportunity random forest

Does fixing bias in tabular ML cost accuracy, and do all mitigation strategies work equally well?

flowchart LR
    A[Tabular dataset] --> B[Bias audit<br/>12 modules]
    B --> C[Reweighting]
    B --> D[Feature suppression]
    B --> E[Post-processing]
    C --> F[Fairness metrics]
    D --> F
    E --> F
    F --> G[Random Forest<br/>accuracy check]

The sharper finding: naive feature suppression, the most intuitive fix, was the least effective of the three, underperforming reweighting on every fairness axis tested. Built with Dr. Chirag Joshi; pending arXiv endorsement.

Contributions to libraries with real production surface area, not toy patches.

• pandas - fixed a negative-slice indexer validation bug in core indexing logic using slice.indices() (PR #66101), with regression tests.
• PyDriller - corrected Commit._stats() to respect the skip_whitespaces flag (PR #320); added a Commit.patch property exposing full unified diffs, closing a long-standing feature request (PR #321).
• PyGithub - added a configurable max_rate_limit_wait cap to GithubRetry, with a new RateLimitExceededExceedsMaxWait exception, replacing unbounded rate-limit stalls (PR #3540).

go raft consensus deterministic simulation testing distributed systems

Can you make a distributed-consensus bug reproduce on demand?

flowchart LR
    A[Raft nodes in Go<br/>leader election + log] --> B[Simulated scheduler<br/>deterministic delivery]
    B --> C[Fault injection<br/>drop / delay / partition]
    C --> D[Seeded runs<br/>replay any failure]
    D --> E[6 passing tests incl.<br/>partition majority/minority]

Every test run is fully deterministic: same seed, same interleaving, same failure. No flaky distributed tests, no "works on my machine" for consensus bugs.

html sheetjs offline-first statistical filtering

Vocational trainee with the APR/SCADA team (June-July 2026). Two shipped tools:

• A fully offline, browser-based CSV/Excel trend-report converter (single HTML file + SheetJS) with a compound tag-editing UI: suffix-diff, character-strip, and per-position removal. Runs on air-gapped plant machines with zero install.
• An outlier-removal and feature-rejection pipeline for Pearson correlation analysis, combining IQR filtering with MAD-based modified z-scores and whole-graph auto-rejection (gap test + leverage test).

clip umap fastapi docker gcp

Can a machine read taste?

flowchart LR
    A[25-round quiz<br/>image comparisons] --> B[CLIP ViT-B/32<br/>embeddings]
    B --> C[Cosine similarity to<br/>16 aesthetic centroids]
    B --> D[3D UMAP projection]
    D --> E[kNN nearest-image<br/>retrieval]
    A --> F[Upload & classify<br/>any photo]

Free-tier GCP VM (964MB RAM) by choice: the constraint is what makes the engineering real. Free HTTPS, a real domain, zero ongoing cost. Full writeup in the repo's README.

• 🎓 Amazon ML Summer School 2026 - selected among the top ~3,000 of 134,000+ applicants (~2.2% selection rate)
• 🏆 Top 1,500 of 100,000+ participants - Google "The Big Code" competitive programming challenge