Anna W.

Own product roadmap for edge AI solutions across industrial and automotive verticals. Lead end-to-end development of scalable AI inference SDKs in coordination with global software and hardware teams. Define KPIs and success metrics for software delivery, partnering with PM, sales, and engineering functions. Manage API performance benchmarks and lead security-focused feature implementations

Built scalable pipelines for data annotation, semantic segmentation, and object tracking. Designed multi-agent trajectory prediction models using LSTM and GNNs. Collaborated with simulation teams to develop synthetic data for rare-edge-case training. Reduced model drift by implementing active learning loops and uncertainty sampling methods

Developed sensor fusion frameworks for combining LiDAR, vision, and IMU inputs in real time. Trained object detection models (YOLOv3, SSD) optimized for edge deployment. Filed two internal IP disclosures on trajectory prediction using dynamic obstacle mapping. Led architecture development for real-time perception and path planning algorithms

Designed firmware modules for SSD and memory controller products. Developed low-level C/C++ code for microcontroller-based systems and led system integration debugging. Created a diagnostic tool that reduced RMA cases by 30% within one year. Coordinated with manufacturing sites in Malaysia and Vietnam to improve test throughput and QA automation

Optimized TensorFlow inference pipelines for Intel CPU deployment. Cleaned and preprocessed over 1TB of video data using automated Python scripts. Evaluated OpenVINO performance for embedded systems in collaboration with the graphics team. Delivered final benchmarking reports with strategic optimization proposals to the R&D division