Position Title:
Lead Data Scientist

Work Arrangement & Location:
100% remote (Denver Based Preferred)
MST time zone working hours. 8-5pm. 40 hours.

Duration:
3.5 Week Contract

Role Overview
Trace3 is hiring a technical lead to own data science delivery for a predictive freezer-failure detection engagement at a Fortune 100 aerospace and defense manufacturer. The engagement turns freezer temperature and equipment diagnostic data into a validated anomaly-detection model, an ROI estimate, and a prioritized rollout roadmap.
This is a hands-on, client-facing lead role. You build the models, shape the scope with the client, set the technical direction the delivery team follows, and keep the delivery artifacts clean enough to hold up under technical scrutiny. Phase 1 is a tightly scoped proof of concept that validates what the existing data can support before any production build.

Key Responsibilities
Lead the Phase 1 data review. Assess the temperature and diagnostic streams for sampling rate, historical coverage, gaps, timestamp alignment, and label availability; publish a clear read on what the data can and cannot support.
Build the anomaly-detection approach. Design and develop failure-detection models from sparse or unlabeled historical sensor data.
Define ground truth with client SMEs. Identify known failure events, near-misses, and maintenance windows, and set a labeling method where formal labels do not exist.
Set measurable acceptance criteria. Agree explicit thresholds for the POC model (detection lead time, false-alarm rate, coverage) before validation, following a tight-scope, defined-tolerance model.
Translate results into a business case. Produce an ROI estimate and a prioritized Phase 2 roadmap the client can sign off on.
Own client communication. Present findings to client stakeholders and Trace3 leadership; maintain documentation and delivery artifacts to a professional-services standard.

Required Qualifications
Experience with developing anomaly detection or predictive maintenance models on sensor or telemetry data.
Experience building models on messy, incomplete, or unlabeled data, including defining labels alongside domain experts.
Advanced degree (MS or PhD) in a quantitative or scientific field.
10+ years applying data science to real problems, with a track record of validated models.
Strong Python for modeling, backed by solid statistics and time-series analysis.
Ability to design end-to-end data pipelines and enforce data integrity and reproducibility through documented SOPs.
Client-facing communication: presenting technical results to non-technical stakeholders and executive or funding audiences.
US Person status (required for ITAR-controlled work).

Preferred Qualifications
Familiarity with industrial and OT data sources: DAQ systems, Kepware/OPC, historians, MQTT.
Experience with lakehouse query engines and dashboarding.
Prior work in regulated or federal environments (defense, CDC/NIH, FDA).
Prior work in regulated or federal environments (defense, CDC/NIH, FDA).