Northwood Space
DSP Engineer
About this role
Northwood seeks a DSP Engineer to develop signal processing algorithms for software-defined radio systems supporting satellite ground stations. You'll own the full lifecycle of modem and beamforming algorithms—from modeling through field validation—working with FPGA teams to deploy critical communication infrastructure globally.
What you'll do
- Design and validate physical layer algorithms for demodulation and beamforming
- Develop floating-point and fixed-point algorithm models in Python and Rust
- Collaborate with FPGA engineers on fixed-point implementation and RTL validation
- Contribute to RF systems analysis and link budgeting for phased array systems
- Own customer waveform requirements and end-to-end delivery
What they're looking for
- Python programming
- Digital communications (modulation, synchronization, demodulation)
- C++ or Rust
- Fixed-point algorithm development
- Signal processing
- RF systems analysis
- FPGA collaboration
- Software-defined radio (SDR) frameworks
Benefits
- Base salary $100K–$170K depending on level
- Work on globally deployed space infrastructure
- Rapid deployment and real-world impact
- Strong venture backing ($100M+)
- Collaborative engineering environment
- Ownership of algorithmic outcomes
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
Northwood Space
Northwood Space builds advanced ground station infrastructure for satellite communications, including phased array antennas and high-speed digital systems deployed globally. The company is hiring engineers across mechanical design, electrical systems, tooling, testing, and infrastructure to support the development and deployment of scalable satellite ground networks.
View all jobs at Northwood SpaceLikely interview questions
- Walk us through a DSP algorithm you've built from concept to deployment—how did you validate it worked end-to-end, and what surprised you?
- Describe your experience converting a floating-point algorithm model to fixed-point. What were the key tradeoffs and how did you verify correctness?