Responsibilities
- Design and develop audio-based detection and classification systems for challenging real-world environments
- Implement robust signal processing pipelines tailored for noisy outdoor conditions
- Build and optimize machine learning models for sound event detection
- Develop low-latency, high-reliability streaming pipelines
- Handle imbalanced and imperfect datasets using augmentation and synthetic data techniques
- Deploy and optimize models on edge hardware platforms (Jetson, Raspberry Pi, etc.)
- Optimize inference performance using ONNX, TensorRT, and OpenVINO
- Develop production-grade Python systems with modular architecture and multiprocessing capabilities
- Ensure system robustness under variable acoustic conditions and hardware constraints
- Collaborate with ML, hardware, and systems engineering teams to deliver integrated solutions
Requirements
- 4+ years in ML / Audio / DSP / Edge AI
- Strong knowledge of audio signal processing (spectrograms, noise reduction, feature extraction)
- Experience working with noisy environments (wind, city, nature)
- Hands-on experience with ML for audio (CNNs, YAMnet, ONNX)
- Proficiency in training on imbalanced datasets and applying augmentation techniques
- Ability to build low-latency streaming pipelines for real-time audio processing
- Experience deploying models on edge devices (Raspberry Pi, Jetson Nano)
- Optimization skills using ONNX, TensorRT, OpenVINO
- Production-level Python engineering experience (clean architecture, multiprocessing, modular pipelines)
- Proven track record of production deployment in real-world scenarios
Nice to Have
- Acoustic domain knowledge (drone frequency ranges, Doppler effect, microphone arrays)
- Sensor fusion experience (audio + video, audio + RF detection)
- Hardware integration skills (GPIO, signal triggering)
