Detection Scientist: Real-World Sensor Analytics

Physics World; United Kingdom; Hybrid; GBP 90,000 - 130,000; Full time

Benefits offered by this job

Rapid career progression

Conference budget

Flexible, hybrid working

NEST pension

Private healthcare

29 days holiday

Job summary

Oxford Dynamics based in the UK is seeking an applied data scientist to turn messy sensor data into trusted detection capabilities. You will work with maritime AIS, signals, and fused feeds, selecting the right techniques and ensuring results are explainable and defensible.

The role emphasizes end-to-end pipeline ownership, from feature engineering to production deployment. A quantitative background helps, with hybrid, flexible working in Oxford/London and strong security clearance requirements.

Qualifications

• Applied data scientist with real datasets, not just toy data.
• Fluent Python and scientific stack; comfortable with at least one DL framework.
• Experience in anomaly detection, time-series, or geospatial analysis.
• Ability to define what good looks like for a detection problem and manage drift.
• Knowledge of production pipelines and reproducible ML/DS workflows.
• Comfort with ambiguity, partial data, and explaining complex results to non-specialists.

Responsibilities

• Build detection capability on real data: signals, AIS data, and multi-source feeds.
• Choose appropriate techniques (ML, deterministic rules, or heuristics) for the problem.
• Engineer features, per-event aggregates, and spatio-temporal patterns.
• Define evaluation metrics and thresholds to balance false alarms.
• Own end-to-end pipelines from idea to production using MLOps.
• Collaborate with engineers and analysts to translate data into mission outcomes.

Skills

Python

NumPy/Pandas

Scikit-learn

TensorFlow/PyTorch

Anomaly detection

Time-series

Geospatial

MLOps

Education

Quantitative degree (physics/math/engineering/CS)

Tools

ArcGIS/ESRI

Palantir Foundry/Gotham

Git

Docker

KServe

Job description

Oxford Dynamics based in the UK is seeking an applied data scientist to turn messy sensor data into trusted detection capabilities. You will work with maritime AIS, signals, and fused feeds, selecting the right techniques and ensuring results are explainable and defensible.

The role emphasizes end-to-end pipeline ownership, from feature engineering to production deployment. A quantitative background helps, with hybrid, flexible working in Oxford/London and strong security clearance requirements.