Instrument

TensorRT

Level: 7/10 | Type: High-Performance Deep Learning Inference SDK | Platform: NVIDIA GPUs, Engine Serialization & FP16/INT8 Quantization

7/10 level — NVIDIA TensorRT inference optimization, FP16/INT8 precision calibration, serialized engine building, and GPU inference acceleration.

Level: 7/10

TensorRTCUDANVIDIA GPUModel OptimizationFP16/INT8

Proficient in NVIDIA TensorRT (7/10) for maximizing GPU throughput and minimizing latency. Converting PyTorch and ONNX models into optimized TensorRT execution engines, layer fusion, kernel auto-tuning, and FP16/INT8 quantization for real-time computer vision inference.

Where this has actually been used

Projects

Zero-Shot & Open-Vocabulary Visual Tracking Suite
NAVIROX — Autonomous Unmanned Vessel CV & 360° LiDAR Fusion POC

Experience

Technical Lead & Solutions Architect at QBS Co.

Products

VERSEYE

Loading content…