I'm a Senior Applied Scientist building production ML systems for diagnostics, reliability, repair recommendations, and hardware/signal decision-making.
Current focus:
- ML for diagnostics, reliability, and repair recommendations on messy operational data
- Ranking, calibration, uncertainty, evaluation, and cost-sensitive decisions
- Multimodal modeling across logs, sensors, signals, and structured telemetry
Background:
- Ph.D. in Electrical Engineering with deep experience in RF modeling, phased arrays, antennas, and EM simulation
- Applied ML experience across production systems, transformers, Bayesian models, computer vision, and signal processing
- Interested in collaborations involving RFID, wireless inference, phased arrays + ML, signal/text fusion, and high-signal evaluation
Python | PyTorch | Hugging Face | scikit-learn | Polars | Jupyter
AWS | SageMaker | Athena | Docker | Git | Linux
MATLAB | HFSS | CST | FEKO | ADS
- Phased-Array-Antenna-Model: Python library and tutorials for phased-array antenna pattern modeling, beamforming, impairments, and visualization.
- EdgeFEM: 3D finite-element electromagnetics solver for RF/mmWave simulation using Nedelec edge elements.
- PyTorch-Vision-Transformers-ViT: Vision Transformer fine-tuning experiments in PyTorch with practical training and evaluation notes.
- Baseball-Pitch-Sequence-Prediction: Sequence-modeling benchmark using LSTM, Transformer, CNN, HMM, Random Forest, Logistic Regression, and AutoGluon with ablations and MLflow tracking.
- Multimodal modeling over text, sensors, signals, and structured telemetry
- RFID and wireless inference problems
- Phased arrays + ML for surrogate modeling, optimization, and generative design
- Evaluation methodology, calibration, ranking, and decision systems under uncertainty
- LinkedIn: https://linkedin.com/in/jhodge007
- GitHub: https://github.com/jman4162
- Email: jah70 at vt dot edu
Duke basketball | hiking | specialty coffee