Job Description
Photon-IV seeks a Staff AI/ML Engineer to lead development of predictive models for hybrid satellite-terrestrial network stability, powering real-time protocol adaptations in defense-grade systems.
About the Role
The Staff AI/ML Engineer will architect and deploy end-to-end ML pipelines from feature engineering on RF telemetry (RSSI, SNR, Doppler) to advanced algorithms and edge-optimized models driving proactive handovers and link optimization in TN/NTN environments.
You'll own the 5-phase workflow: data prep, feature engineering (temporal/spectral/domain), model design (multi-task loss: MSE/BCE/Huber), training ), and deployment (INT8 quantization, TensorRT for 50ms edge latency).
Key Responsibilities
Design TFT architectures with variable selection networks, multi-head attention, and gated residuals for multi-horizon link stability forecasting (target: 87% accuracy).
Engineer features from baseband KPIs (Doppler shift, path loss, scintillation) using FFT, SHAP, and recursive elimination to 32-64 dims.
Implement training pipelines with early stopping, gradient clipping, and noise augmentation; optimize for cloud
Deploy via ONNX/TensorRT for real-time inference, integrating predictions (stability score 0-1, degradation prob, time-to-failure) into PHY/MAC logic.
Collaborate with wireless engineers to close AI-protocol loops, validating in hybrid sims with mobility/Doppler/propagation delays.
Required Qualifications
5+ years production ML experience; MSc/PhD in CS/EE/AI or equivalent.
Expertise in time-series forecasting (TFT, LSTM, TCN) and PyTorch/TensorFlow; strong Python (PySpark, NumPy, scikit-learn).
Proven deployment: quantization, ONNX, TensorRT/NVIDIA edge optimization.
Familiarity with wireless metrics (RSSI/SNR/BER, jitter/latency) or ability to rapidly learn NTN standards (3GPP, IEEE 802.16).
Preferred Skills
Experience with RF/satellite data: spectral analysis, Doppler compensation, atmospheric effects.
Full-stack ML: RAG/LLMs, transfer learning, explainability (SHAP/UMAP).
C++/Rust for performance-critical inference; NVIDIA Inception/GPU cloud scaling.
If you've shipped AI models turning signal chaos into resilient connectivity or something similar, apply with resume, GitHub, and note on hybrid TN/NTN ML projects. photon-iv.com/careers.
Job Type: Full-time
Pay: $50.00-$65.00 per hour
Expected hours: 37.5 per week
Experience:
wireless: 5 years (preferred)
Work Location: Remote
