Song WANG

I was a Huawei Level-17 Senior Engineer (A) and hold the qualification for Level-18 Principle Engineer. My research and development focus was on AI algorithms and applications. First, I spearheaded the development of a hardware-in-the-loop autonomous driving simulation system. I constructed the hardware infrastructure, integrating an Ubuntu server running the Carla simulator and an embedded Linux platform via ROS. Furthermore, I refactored an open-source autonomous driving algorithm called LAV using C++ for edge deployment. The complete simulation system was packaged using Docker and delivered on time, with all performance expectations been met, earning me a High Quality & Efficiency Award. Second, I played a leading role in a project developing a cascade of algorithms to efficiently partition neural networks for distributed deployment across multi-chip systems, maximizing parallel computing performance. I spearheaded the development of a topological network condensation algorithm and an ONNX-based partition algorithm. In addition, I also led the final Docker-based integration and delivery of our team’s collective work. Ultimately, I was granted an Innovation Breakthrough Award. Third, I developed reinforcement learning and imitation learning algorithms, such as Soft Actor-Critic and ACT, for end-to-end control of a robot arm using visual inputs. These algorithms successfully executed in-simulator tasks, including 1) pushing cubes to target zones and 2) grasping and plugging in a connector. Fourth, I conducted exploratory research into generative AI technologies, focusing on 3D generative models and Vision-Language-Action (VLA) frameworks. I successfully developed an image-to-3D generative model, utilizing Diffusion Transformers that were trained with flow matching techniques. I also developed a VLA model for autonomous driving, featuring a semantic-level tokenizer for visual information encoding and a Diffusion Head for generating future waypoints.

Telecommunication base stations need capability to allocate different resources across 2G, 3G, 4G and 5G networks. Our team’s task was to develop and maintain a profile creator that generates configuration profiles for Nokia base stations. My role focused on implementing new features to support emerging 5G technologies, alongside routine maintenance and bug fixes. I also served as a key point of contact for cross-border collaboration with colleagues in Poland. Prior to my resignation, I was honored with the TC Recognition Award from Nokia Hangzhou Technology Center.

The company specializes in producing assembled power battery packs. My mission was to develop a lithium battery management system (BMS) based on HCS12X and MCS-51 microcontrollers. I designed hardware schematics and PCBs, integrating a diverse array of ICs and sensors for control and power management. I also implemented system firmware, including an integrated user controller featuring an OLED display for real-time monitoring, multiple control buttons, and a configuration menu. In addition, I set up field test environment and conducted unit tests and system tests. The complete product design has been documented and submitted as a Chinese invention patent.

My study field was Intelligent Systems Automation. I learned industrial engineering and conducted research in areas of haptic sensing, robotics and machine learning. My main research topics focused on developing a human balance sensor and another skin sensor based on the principle of frustrated total internal reflection, along with the corresponding machine learning and computer vision algorithms for data processing, which led to four peer-reviewed paper publications and one U.S. invention patent. Furthermore, I proposed an innovative Generative Adversarial Tri-model (GAT) method, which can integrate data-driven neural networks with analytical knowledge, which led to three peer-reviewed paper publications. In addition, I got experience being a teaching assistant for lab tutorials about Robotics and ROS. I totally published 7 peer-reviewed papers and 2 U.S. patents. I was an awardee of the HKU Postgraduate Scholarship and the Arthur and Louise May Memorial Fund Scholarship (twice).

My specialty was General and Fundamental Mechanics. I learned advanced mechanics and performed research investigating elastic vibrations. My main reasearch topic focused on revealing the transient dynamic behavior of elastomers such as bars, beams and rings during collision processes, which led to two peer-reviewed paper publications. My collaborative research topic focused on designing a vibration energy harvesting device with a bistable structure optimized for low-frequency environments, which led to another peer-reviewed paper publication. I was awarded the Ying Huaiqiao Scholarship and "Excellent Postgraduate Student Award" of Zhejiang University. My thesis won the "Excellent Master's Thesis Award" of Zhejiang Province.

My major was Automation. I learned mathematics, physics, electrics and electronics, control theory and computer programming. I carried out projects about embedded systems design. I developed PID control algorithms for inverted pendulum and two-wheel self-balancing vehicle based on STM32 microcontrollers. I also implemented wired communications through UART, SPI, I2C protocols and wireless communications using both Bluetooth module and NRF24L01 transceiver. I was awarded the National Scholarship twice. In 2012, I led a team to win a National First Prize in the prestigious National College Student Intelligent Car Competition by developing an MCU-controlled line-following wheeled robot.