Teams at VinFast Auto Ltd.

Develop, deploy, and optimize CNN/DNN and computer-vision models on embedded SoCs and hardware accelerators. Convert models across PyTorch, TensorFlow, ONNX, and target runtimes; optimize quantization, graph execution, memory, latency, throughput, power, and accuracy. Integrate and debug end-to-end camera-to-application inference pipelines in embedded Linux using C/C++ and Python across heterogeneous CPU, GPU, DSP, and NPU architectures.

Leads VinFast’s passive safety and crashworthiness function from vehicle concept through SOP. Owns crash architecture, occupant protection, restraint development, CAE-to-test correlation, crash-test root-cause analysis, NCAP and regulatory readiness, and EV post-crash integrity. Drives structural, restraint, packaging, material, and calibration countermeasures while balancing safety, mass, cost, timing, and manufacturability. Builds the Passive Safety organization, establishes engineering standards, and serves as the senior technical escalation point.

Designs and executes manual and automated tests for AI systems, including data, functional, model, security, safety, performance, regression, and system testing. Evaluates hallucination, bias, fairness, and model accuracy; develops test scripts, reports defects, collaborates with technical stakeholders, and improves QA processes for reliable AI product releases.