Production Deployment: Engineered and deployed robust deep learning solutions for industrial defect detection, handling real-time inference on edge devices with strict latency constraints (<5 ms). Performance Optimization: Accelerated legacy Python pipelines by rewriting code in C++, achieving a great speedup in batch processing and enabling high-speed manufacturing throughput. ML Operations (MLOps): Standardized the model delivery process using Docker, ensuring reproducibility across development and production environments at client sites. High-Performance Integration: Architected a unified C++ Hardware Abstraction Layer to interface with heterogeneous industrial devices (Teledyne DALSA, Alkeria cameras, Keyence profilometers), decoupling hardware logic from the core application. Edge Computing & Deployment: Led the containerization (Docker) and deployment of quantised Neural Networks (ONNX, TensorRT) on resource-constrained edge devices (NVIDIA Jetson), balancing accuracy with strict

Software Engineer — High-Performance Systems & C++ at T3LAB

April 1, 2024 - Present

Production Deployment: Engineered and deployed robust deep learning solutions for industrial defect detection, handling real-time inference on edge devices with latency under 5 ms. Performance Optimization: Rewrote legacy Python pipelines in C++, achieving significant speedups in batch processing and enabling high-speed manufacturing throughput. ML Operations (MLOps): Standardized the model delivery process using Docker for reproducibility across development and production environments at client sites. High-Performance Integration: Architected a unified C++ Hardware Abstraction Layer to interface with Teledyne DALSA, Alkeria cameras, and Keyence profilometers, decoupling hardware logic from the core application. Edge Computing & Deployment: Led containerization and deployment of quantised Neural Networks (ONNX, TensorRT) on NVIDIA Jetson devices with strict latency requirements. Internal Tooling: Developed diagnostic desktop tools (Qt/C++) used by field engineers to debug sensor data s