I build C++ and Python tools for Linux evidence, repository hygiene, and software supply-chain review. My focus is systems programming and defensive security: making inputs, decisions, and output contracts inspectable.
Start with the code below. The portfolio website adds selected engineering cases and a five-minute review route.
| Project | Engineering focus | Reproduce / inspect | Tagged release |
|---|---|---|---|
| LogLens | C++20 auth-log parsing, coverage telemetry, detection episodes, Markdown/JSON reports | Reviewer path | v0.6.0 |
| telemetry-lab | Five local detection/investigation demos, schema contracts, and run manifests | Operator reproduction | v1.2 |
| repo-sentinel-lite | Heuristic repository scans, redaction, reviewed baselines, and changed-file gates | Consumer setup | v0.8.1 |
| scientific-computing-toolkit | SBOM/dependency diffing with JSON, Markdown, SARIF, local policy, and optional provenance | SBOM reviewer path | v1.1.0 |
Supporting work:
- systems-foundations: four saved-evidence Linux mini-labs, including a process-diff-to-JSONL adapter. Process workflow and tests · tagged v0.3.0.
- sec-writeups-public: security pattern cards with supporting lab notes and Python publication checks.
Tag links describe released versions. The source and tests linked on main can include subsequent work. Detection demos remain local labs; scanner and policy outputs support review without certifying that a system or package is safe.
Saved process evidence → telemetry aggregation connects systems-foundations to telemetry-lab at pinned public commits. PR #96 merged after v1.2. The article reports successful ingestion and aggregation with zero alerts; it does not establish detection efficacy. Code, fixtures, and producer goldens are public; full execution receipts remain privately retained.
Authored PRs, with status checked 20 October 2026. Contribution scope is the linked diff; the upstream projects retain their own authorship and review process.
- Open — microsoft/sarif-tools #115: preserve the full result description after a rule ID, with regression tests.
- Open — microsoft/sarif-tools #117: contextual filter-configuration diagnostics and advisory zero-match warnings.
- Open — ossf/pvtr-github-repo-scanner #513:
return
NeedsReviewwhen check-suite/run evidence is incomplete, with Go regressions.
- LogLens: parser uncertainty as evidence
- telemetry-lab: evidence pipeline contracts
- systems-foundations: Linux notes
Public work uses sanitized fixtures, explicit input/output contracts, and reproduction commands. AI assistance may be used for implementation and drafting; tests, artifact inspection, and human review determine what is published.
Focused issues are welcome on parser coverage, schema compatibility, provenance, or policy decisions. Include the exact command and a sanitized example in the relevant repository.