传感器算法实习生-上海

职位描述

Responsibilities - Prototype and implement DSP algorithms, including filtering, time- and frequency-domain analysis, and feature extraction, for sensor applications under the guidance of the team. - Analyze raw sensor data and contribute practical solutions for de-noising, detection, and parameter estimation against project milestones. Support hardware engineers on the sensor signal chain, including raw-signal processing, noise analysis, and calibration checks. - Help translate a feature definition into testable use cases, then implement and evaluate the corresponding algorithm. - Model and simulate algorithm designs and report performance against agreed metrics. - Validate algorithms with Monte Carlo simulation and recorded or field data. Compute and track detection rate, false-alarm rate, SNR, and latency. - Apply machine-learning methods where they are appropriate, with mentoring on data collection, training, and validation. Use MATLAB, Python, or C++ to analyze data, implement algorithms, and document results. - Work with platform software engineers to understand embedded constraints such as fixed-point arithmetic, memory, and real-time execution, and support a prototype implementation where needed. Collaborate with mechanical, hardware, and software engineers. Clearly document assumptions, results, and open issues. - Debug mismatches between simulation and measured data, and propose the next experiment or code change. Requirements - Current Master’s or PhD student, or a strong final-year Bachelor’s student, in Electrical Engineering, Computer Engineering, Physics, or a related field. - Solid DSP coursework or project work covering sampling theory, the Z-transform, windowing, spectral estimation, statistical signal processing, probability, and power spectral density estimation. - Hands-on implementation, not only library calls, of filtering and frequency-domain analysis such as FFT or STFT. Kalman, EKF, UKF, or other tracking experience is a plus. - Ability to compute and explain SNR, detection rate, false-alarm rate, and error or latency, using Monte Carlo simulation or equivalent experiment design from coursework or research. - Practical experience in MATLAB and/or Python for scientific analysis and algorithm development. C or C++ is a plus. - Experience processing raw or near-raw sensor data, such as ADC samples, IQ data, array snapshots, or correlator output, from coursework, a thesis, or a research project. - Clear written and verbal communication, and the ability to explain a noise model, a processing chain, and a result curve. Self-driven execution on a scoped problem, with regular check-ins rather than unsupervised ownership of a product. Preferred - Sensor calibration or gain, phase, frequency-response, or bias analysis, and any hardware/software co-debug experience. Exposure to automotive or robotics sensors such as radar, ultrasonic, IMU, or microphone arrays. - A prototype that ran on an embedded or resource-constrained platform, including fixed-point or memory-aware implementation. - Coursework or project experience in machine learning for signal or sensor data. This is not a substitute for DSP fundamentals. - Experience plotting results and running basic statistical comparisons.

任职要求

• Python
• Go
• C++