Machine Learning Engineer | Geometric AI | Handshake

[Note: This position is contingent upon contract award.]

Job Title: Machine Learning Engineer

About the Role:

We are seeking a brilliant, physics-minded Machine Learning Engineer to help us redefine the state-of-the-art in target detection and classification.

We are building architectures that natively ingest, process, and learn from raw wave physics. You will be at the forefront of developing Physics-Informed Deep Learning Architectures to process complex-valued phase history and element-level array timeseries data.

If you understand the Data Processing Inequality, if you believe that every dB of signal matters, and if you know how to map raw geometric physics into neural network architectures, we want you on our team.

What You Will Do:

Architect Advanced Neural Networks: Design and train novel deep learning models (such as spatial-temporal or spectral architectures) capable of processing raw physical features without breaking phase continuity.

Process Raw Waveform Data: Work directly with uncompressed, terabyte-scale phase history data and raw multi-channel spatial array timeseries.

Implement Physics-Informed Priors: Integrate exact wave physics—such as geometric propagation delays, spatial coherence, and wavefront curvature—directly into neural network formulations to create optimal, learned spatial filters.

Build Bulletproof Data Pipelines: Write elegant, memory-safe, highly vectorized Python (NumPy/SciPy) pipelines to extract multi-dimensional features (energy, phase variance, spatial gradients) from massive datasets without triggering out-of-memory errors.

Optimize Sampling Strategies: Develop intelligent, targeted sampling algorithms to build highly representative statistical mosaics of physical environments.

The Tech Stack You Will Use:

Deep Learning: PyTorch.

Data & ML: NumPy, SciPy, Scikit-Learn.

Signal Processing: Fast Fourier Transforms (FFTs/IFFTs), windowing, beamforming, and complex array math.

Environment: Linux, Google Cloud Platform (GCP) heavy-compute instances.

What You Need to Have (Requirements):

Education & Experience: New college graduates are highly encouraged to apply! A Master’s degree in Computer Science, Electrical Engineering, Physics, Applied Mathematics, or a closely related quantitative field is highly preferred.

Math & Physics Foundation: A strong academic or intuitive understanding of wave interference, phase coherence, Doppler effects, and signal-to-noise ratio mechanics.

Neural Network Knowledge: Academic or project-based experience building and training deep learning models, specifically dealing with spatial arrays, time-series, or digital signal processing.

Strong Python Programming: You write clean, vectorized Python code and have experience managing memory efficiently when dealing with large datasets.

Domain Exposure: Familiarity with Digital Signal Processing (DSP), RF Engineering, or general wave propagation.

Ability to obtain or maintain a U.S. Security Clearance.

Bonus Points (Standout Qualifications):

Experience with spectral processing or advanced representation learning.

Interest in or experience with government/DoD research proposals (SBIRs) and working with restricted datasets.

Why Join Us?

You won't be fine-tuning large language models or building standard CNNs for basic image classification. You will be solving fundamentally hard physics problems using bleeding-edge mathematics, directly translating theoretical limits into operational realities. You will have the compute budget to do it right, and the mandate to capture every last decibel of signal.