AstraSense Diagnostics and Prognostics Engineer

Location: Los Angeles, CA or Louisville, KY

About the work

Develop AstraSense health models for electric propulsion and related aircraft subsystems. Combine signal processing, physics-based models and machine learning to detect anomalies, investigate faults and estimate degradation. Make health outputs useful to engineering, maintenance and mission decisions while stating their uncertainty and operating limits.

What you will do

• Build versioned telemetry and training-data pipelines for motors, inverters, high-voltage cabling and battery packs, starting with a defined subsystem and an agreed evaluation dataset.
• Extract FFT, envelope, RMS and harmonic-distortion features from electrical, vibration and acoustic signals; evaluate high-rate channels sampled at 20 kHz or higher where required by the agreed subsystem specification.
• Develop anomaly-detection, fault-classification and remaining-useful-life approaches. Establish simple baselines, account for changing operating conditions and quantify false alarms, missed events and uncertainty.
• Generate physics-informed synthetic fault signatures for cases such as insulation degradation, harmonic distortion and bearing wear. Label simulated and measured data clearly, document assumptions and avoid treating synthetic coverage as physical validation.
• Prevent train/test leakage across runs and assets; maintain reproducible experiments, data provenance, model versions and test reports.
• Work with embedded engineers on compression, quantization and target-hardware evaluation toward the specified 100 ms capture-to-detection budget. Report measured performance and limitations instead of assuming the target is met.
• Compare results with available bench and, when available, approved flight datasets; define health/capability outputs and interfaces with systems engineers.

What you bring

• Practical Python and experience with time-series data, statistics, signal processing or reliability modeling.
• Ability to design reproducible experiments, prevent data leakage and explain model limitations.
• Engineering, applied mathematics, computer science or equivalent practical experience.
• Familiarity with scientific Python and an ability to select suitable tools, such as NumPy, SciPy, pandas, scikit-learn, PyTorch or TensorFlow.

Useful additional experience

Predictive maintenance, batteries, motors, power electronics, aircraft systems, MATLAB/Simulink, remaining-useful-life estimation, condition monitoring or resource-constrained model optimization.

First 90 days: scope to agree

Agree a scoped first-90-day objective: a reproducible anomaly-detection baseline for one subsystem on available bench data, plus a reusable synthetic fault dataset. Record assumptions, false-alarm/missed-event results, uncertainty and an embedded integration plan. The scope and acceptance criteria will be agreed against available data and hardware.

About AstraQua

AstraQua is developing a modular autonomy architecture with AQFM as its shared intelligence foundation. Specialized agents contribute perception, reasoning, health and coordination. AstraSense is the health-intelligence agent, with electrified aviation as an initial research application; it connects condition and remaining capability with mission decisions. Our approach to safe AI emphasizes defined operating boundaries, deterministic safety controls, operator authority and behavior designed for verification and validation.

How to apply

Email your resume, job title and preferred location (Los Angeles, CA or Louisville, KY) to careers@astraqua.com. Include a nonconfidential project example.