Job Type
Full-time
Description
Type: Full-Time(W2) On-site/Hybrid, Arlington, VA (Remote option available for the right candidate)
In this role, you will design, prototype, and validate novel AI/ML components-such as neural receivers, neural beamforming, neural scheduling, digital twin, and ISAC (Integrated Sensing and Communications)-that outperform traditional signal processing methods. You will work at the cutting edge of 6G innovation, taking concepts from mathematical intuition to simulation (e.g. NVIDIA Sionna) and real-time implementation.
What You'll be Doing - Applied AI Research: Design and train modern deep learning models (Transformers, Vision architectures, etc.) to solve complex physical layer problems, including channel estimation, MIMO detection, and beam management
- Simulation & Validation: Build high-fidelity link-level simulations using NVIDIA Sionna and ray-tracing to train, test, and benchmark AI models against legacy 5G baselines
- Prototyping & Deployment: Transition research models into deployable "dApps" for the Distributed Unit (DU), optimizing inference for latency and compute efficiency on NVIDIA GPUs
- New Capabilities: Explore emerging AI-RAN frontiers such as Integrated Sensing and Communications (ISAC), neural scheduling, and channel digital twins
- Innovation & IPR: Drive technical innovation by authoring invention disclosures, filing patents, and generating technical reports to support our standardization team in 3GPP and O-RAN Alliance contributions
- Data Engineering: Architect data pipelines for generating synthetic training datasets and developing "Sim-to-Real" transfer techniques to ensure robust performance in real-world networks
Required Qualifications- Education: Ph.D. or Master's in Computer Science, Electrical Engineering, or Applied Mathematics with a focus on Deep Learning and/or Communications Systems
- AI/ML Expertise: 3+ years of experience designing and training deep neural networks from scratch. Strong grasp of modern architectures and optimization techniques
- Applied Signal Processing: Experience applying machine learning to real-time time-series data, signal processing, or physics-based problems (Audio, RF, or similar domains)
- Research to Code: Proven ability to read academic papers and implement their methods in robust Python code
- Simulation Skills: Experience with differentiable simulation or digital twins (e.g., Sionna, JAX-based physics sims)
Preferred Qualifications- Wireless Knowledge: Understanding of wireless fundamentals (OFDM, MIMO, IQ data) is highly helpful, though we prioritize strong ML intuition over pure communication theory
- Performance Optimization: Experience with model quantization (FP16/INT8), pruning, or using TensorRT for real-time inference
- Standardization Support: Experience writing technical whitepapers or supporting patent filings in a research environment
- C++ Integration: Ability to write C++ bindings or integrate Python models into C++, SIMD, and Cuda production pipelines
