Open source engineer building data platforms, developer tooling, and technical writing around Rust, Python, Go, and modern analytics systems.
These are my projects, contributions, and writings. For more about me, use the links below.
Website | Blog | LinkedIn | Email
A curated view of the systems I actively build. Related papers, integrations, datasets, and distribution repositories are grouped with their parent project.
| Project | Why it matters |
|---|---|
| dataprof | Bounded-memory Rust/Python profiling for files, DataFrames, and Arrow streams, with explicit assessment scope and comparable quality reports. Applied in Metric Evidence, a Power BI report that shows data evidence next to each metric. |
| arrow-abi-doctor | Differential testing of the Arrow C Data and C Stream interfaces across real consumers, in isolated processes. |
| Tabular Evolution Corpus | Versioned Parquet scenarios of datasets changing shape, with machine-checked manifests, also on Hugging Face. |
| Nephtys | A low-footprint, multi-protocol connector for durable NATS JetStream pipelines, backed by an IEEE UIC 2026 evaluation. |
More selected work — open data, knowledge graphs, lakehouse, applied research, and systems work
| Project | Focus |
|---|---|
| Ceres + Ares | v0.7.0 / v0.4.0 — reproducible open-data harvesting and schema-driven extraction. Ceres powers a public open-data index and a Databricks analytics pipeline. |
| GitNodes | v0.2.0 — a Git-native, typed knowledge graph that engineering teams and AI agents can explore through the web UI or MCP. |
| dbt-isrm | New — a release matrix for dbt's Information Schema across releases and execution stages. |
| dlt + dbt on Databricks | Active — runnable ingestion, transformation, orchestration, and quality-gate examples validated against Databricks and in DuckDB CI. |
| OCCAS | v0.1.0 research release — a provenance-backed PX4/ERA5 agreement study; external wind estimates were too imprecise to substitute for onboard estimates under the study’s criterion. |
| Fantabuddy | v0.4.0 data product — a local DuckDB history of official player lists, offline Classic/Mantra auction reports, and verifiable Parquet corpus exports. |
| DCE | Pre-release — a Rust-native engine for defining, validating, and enforcing data contracts. |
| Zero Grappler | Prototype — a no_std, allocation-free async pipeline for embedded sensor, DSP, and inference workloads. |
| IcebergSharp | In development — .NET Iceberg metadata, Avro manifests, and REST catalog access; scan planning and Parquet data reads remain on the roadmap. |
| Lares | In use — a small self-hosted home stack built around failure semantics: digest-pinned images, application-consistent SQLite snapshots to encrypted offsite backup, and a scripted bare-metal rebuild. |
Browse the full project archive or all repositories.
Contributions to the broader open source ecosystem beyond my own repositories.
This section is updated automatically via GitHub Actions.
63 merged PRs across 29 upstream projects · 219.4k combined stars
