National Health Workforce Registry
Volunteer project, public health data
A registry, API and public dashboard for a national health regulator. 8,627 facilities catalogued; location matches await field verification.
Projects in reporting, Power BI and analysis.
Public-interest work
Supporter Survey to Work Plan
Volunteer project, confidential survey data
89 supporter responses for a one-person climate nonprofit, coded into three recommendations. No impact was measured.
Fair Lending Lab
Statistical analysis, public data
The raw denial-rate gap of 10.7 percentage points fell to 4.73 after the available financial controls.
Water System Risk & Funding Priority Index
Risk modeling, Ohio public data
0.740 ROC AUC against 0.734 for counting prior violations, so the composite did not earn its complexity.
Forecasting, optimization and recommendation
Grid Intelligence
Forecasting dashboard, public data
5.49 percent MAPE against 8.30 percent for the EIA day-ahead forecast, lower in 35 of 48 shared balancing authorities.
Dispatch Optimizer
Decision optimization, synthetic benchmark
Greedy plus 2-opt beat naive on 99% of randomized scenarios and lost on none.
The interface image shows one seeded synthetic day; the benchmark covers 200 scenarios.
arXiv Recommender
Recommendation systems, public research data
The hybrid reached MAP@10 0.199 against 0.149 for TF-IDF on 2,000 held-out seeds. MiniLM alone lost to TF-IDF.
Tools and applied AI
Cellwright
Open source tool, Excel VBA migration
All 30 third party workbooks in a real corpus inspected on Windows and Ubuntu, with no macro run.
Manufacturing Intelligence Platform
Decision dashboard, synthetic data
The pipeline recovered the signals seeded into 35.3 million synthetic records.
Automotive Analyst
Applied AI, synthetic manufacturing data
After I fixed a bypass the attack suite found, the guardrail blocked 43 of 43 attacks and passed 16 of 16 real queries.
Black Box AI
Applied AI, public data
Hybrid retrieval reached 0.902 MRR against 0.853 for semantic search on 7,462 NTSB reports, not enough to switch.
Evaluation appendix: 21 measured comparisons
Every figure in these case studies that was measured against a baseline. The units differ and are not comparable to each other; the only thing all 21 share is which way the result came out.
- For the build
- 12
- Against it
- 8
- No separation
- 1
| Project | Compared | Figure | Unit | Direction |
|---|---|---|---|---|
| arXiv Recommender | Hybrid blend against TF-IDF | 0.199 | MAP@10 | For |
| arXiv Recommender | Hybrid over TF-IDF on MAP@10 | 1.34× | Ratio | For |
| arXiv Recommender | MiniLM neural tower against TF-IDF | 0.121 vs 0.149 | MAP@10 | Against |
| Automotive Analyst | Adversarial SQL suite against the guardrail | 43 / 43 | Attacks blocked | For |
| Automotive Analyst | Legitimate analytical queries against the guardrail | 16 / 16 | Queries accepted | For |
| Automotive Analyst | Quoted identifiers against my own allow-list | 1 bypass | Bypasses found | Against |
| Black Box AI | Hybrid retrieval against semantic retrieval | 0.902 vs 0.853 | MRR | No separation |
| Dispatch Optimizer | Greedy plus 2-opt against the naive baseline | 12 fewer | SLA breaches per day | For |
| Dispatch Optimizer | Greedy plus 2-opt against the naive baseline | 3.35 hrs | Overtime per day | For |
| Dispatch Optimizer | Cold CP-SAT at the shipped 8s budget against greedy plus 2-opt | 6 behind | SLA breaches per day | Against |
| Dispatch Optimizer | Warm-started CP-SAT at 8s against greedy plus 2-opt | Parity | SLA breaches per day | Against |
| Dispatch Optimizer | The naive baseline measured against itself | 45.6% | Deadlines missed | Against |
| Grid Intelligence | Model against the published EIA day-ahead forecast | 5.49% vs 8.30% | MAPE | For |
| Grid Intelligence | Model against the published EIA day-ahead forecast | +10.0% | RMSE skill | For |
| Grid Intelligence | Model against EIA, counted per authority | 35 / 48 | Balancing authorities | For |
| Manufacturing Intelligence Platform | Injected process events against what the pipeline recovered | Round trip | Signal recovery | For |
| Manufacturing Intelligence Platform | Propagation join against assuming the origin station | 69% | Defects downstream | For |
| Manufacturing Intelligence Platform | Rediscovered step change against the seeded one | 43% | Defect reduction | For |
| Water System Risk & Funding Priority Index | Weighted index against counting prior violations | 0.740 vs 0.734 | ROC AUC | Against |
| Water System Risk & Funding Priority Index | Compliance component alone against the full index | 0.781 | ROC AUC | Against |
| Water System Risk & Funding Priority Index | Index stripped of compliance against chance | 0.447 | ROC AUC | Against |