Analysis / Full-time
Senior Data Scientist, Detection Statistics
Develop statistical methods that separate real persistent signals from the enormous volume of noise the Laboratories record.
The Laboratories receive more candidate reductions than the research staff can manually inspect. This role protects the review standard before volume turns into false confidence.
Resume or CV
Notebook, methods note, or statistical analysis sample
Short statement on rare-event false-alarm control
Candidate exercise
A review of a small candidate-scoring notebook with emphasis on reproducibility and failure cases.
Analysis recruiting desk / FAL-PSC-26-011
Full-time exempt staff appointment
Continuing employee benefit tier
REC-PSC-26-011-ANA
FAL-PSC-26-011-public-notice-04
July 5, 2026
Search committee
- 01Dr. Wen Li
- 02Analysis recruiting desk
- 03Scientific Review Committee delegate
Recruiting active file plus finalist review record
How this packet is evaluated
Analysis
Finalist review covers reproducible analysis habits, model-audit practice, and access to non-public scoring records.
Complete packets are logged under REC-PSC-26-011-ANA, separated from public correspondence, and assigned to the Analysis recruiting desk for first read.
Any exercise is bounded, redacted, and evaluated only for hiring. It is not added to the catalog, public systems, grant files, or laboratory work product.
| Criterion | Weight | Evidence reviewed |
|---|---|---|
| False-alarm control | 35% | Understands base rates, nulls, thresholds, drift, and model overconfidence. |
| Reproducibility | 25% | Leaves notebooks, methods, and replay paths another reviewer can follow. |
| Operational fit | 20% | Works comfortably with long-running queues and imperfect historical records. |
| Communication | 20% | Explains model limits without jargon or false certainty. |
- 01Reference check after candidate notice
- 02Work-location and access-boundary confirmation
- 03Records-handling and confidentiality review
- 04Program-office recommendation before offer
Candidate review docket REC-PSC-26-011-ANA; public posting version FAL-PSC-26-011-public-notice-04.
Responsibilities
- Maintain false-alarm-rate frameworks and coincidence-significance tests
- Triage candidate detections flagged by the Distributed Analysis Initiative
- Build reproducible review notebooks for public and internal catalog entries
- Partner with Stratis engineers on scoring, drift detection, and model audit trails
Qualifications
- Graduate degree in statistics, physics, applied mathematics, or related quantitative field
- Strong background in signal processing, Bayesian or frequentist rare-event methods
- Fluency in Python or another scientific computing environment
- Experience explaining uncertainty to non-specialist decision makers
Preferred experience
- Experience with astronomy, geophysics, or low-frequency acoustic data
- Model governance or ML audit experience
What good work looks like
- False-alarm estimates are reproducible from public and internal notebooks
- Distributed-analysis outputs reach researchers with clear confidence labels
- New scoring methods include failure cases before they are used in review
Continuing staff appointment
Two station residencies per year for audit and catalog review
Limited, usually for annual methods review
Data-access review and model-governance attestation required
Hiring process
- 1Applied statistics screen
- 2Notebook review
- 3Panel with research and Stratis liaisons
Recruiting acknowledges complete packets within five business days and routes them to the program office named on the requisition.
Access, relocation, and accommodations
No station acknowledgement required at application stage
Relocation support is reviewed after finalist stage for continuing Calder Ridge appointments.
Candidates may request accommodation support through recruiting before any timed exercise or station visit.
Analysis recruiting desk, FAL-PSC-26-011
- Recruiting will not request references until the candidate is notified.
- Finalist access checks are started only after a program panel recommends continued review.
- Candidate materials are retained under the recruiting records schedule and are not used as public work product.
Candidate materials remain in the recruiting record and are not used as laboratory work product, public samples, or source material for the catalog.
Candidates may request accommodations before any timed exercise, station visit, or access-review step by contacting the recruiting desk listed on the requisition.
FAL evaluates candidates without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, veteran status, age, or any other protected status.
Apply for this position
Applications are read by the relevant program. Include links to your CV, publications, or portfolio where relevant. For station roles, use the message field to note any field, access-controlled, archive, or remote-site experience that would help the review panel.