Amentum

Job Summary
Build, train, and deploy large-scale, self-supervised "foundation" models that learn rich representations of time series, sequential sensor data in addition to textual and vision data, to be fine-tuned for tasks such as predictive maintenance, forecasting, classification, or multi-modal sensor fusion for industrial and scientific applications.

Duties

• Time Series & Sequential Data - processing, augmentation, feature engineering for financial, industrial, IoT, medical, or other sensor streams.

• Sensor Data Analysis - expertise with diverse sensor modalities, sampling rates, synchronization, and real-world artifact handling.

• Multi-Modality Learning - integrating heterogeneous data types into robust deep learning architectures; cross-modal representation learning.

• Self-supervised and Semi-supervised Learning - time series foundation models, masked modeling, contrastive methods, temporal predictive coding, multimodal alignment and fusion.

• Model Architectures - sequence models (RNNs, GRU/LSTM, TCN), 1D/2D/3D CNNs, Transformers (BERT, ViT, TimeSFormer), graph neural networks, diffusion/generative models, multi-modal/fusion encoders.

• Transfer Learning & Fine-Tuning at Scale - prompt/adapter-based strategies, temporal domain adaptation, few-shot learning for specialized tasks.

Software & Infrastucture

• Programming: expert Python (NumPy, SciPy, Pandas), C++/CUDA for custom kernels and high-performance preprocessing.

• Deep Learning Frameworks: PyTorch (Lightning, Distributed), TensorFlow/Keras, JAX/Flax.

• Large-scale Training: multi-GPU, multi-node clusters, mixed-precision, ZeRO optimization, scalable data loaders for long sequences.

• Data Engineering: robust pipelines for ingesting, cleaning, segmenting, and aligning large-scale, time-synchronized multi-sensor datasets.

• Linear Algebra, Probability & Statistics, Optimization (stochastic, convex/non-convex, Bayesian).

• Signal Processing: Fourier/wavelet analysis, filters (Kalman, Savitzky–Golay), resampling, noise modeling.

• Numerical Methods: ODE/PDE solvers, inverse problems, regularization, time-frequency methods for complex systems.

Job Type: Full-time

Pay: $130,000.00 - $200,000.00 per year

Work Location: In person