Responsibilities
EfficientDet-Lite ๊ธฐ๋ฐ ์จ๋๋ฐ์ด์ค ๊ฐ์ฒด ์ธ์ ๋ชจ๋ธ ์ ์ด ํ์ต ๋ฐ ๊ณ ๋ํ
Fine-tune and enhance on-device object detection models based on EfficientDet-Lite through transfer learning.
๋น ๋ฅด๊ณ ์ ํํ ์ฌ๋ฌผ ์ธ์์ ์ง์คํ ์ฑ๋ฅ ํ์ธ ํ๋
Optimize model performance with a focus on fast and accurate object detection.
์ง ์ ์ฌ๋ฌผ ํ๊ฒ ์ปค์คํ ๋ฐ์ดํฐ์ ์์ง ํ์ดํ๋ผ์ธ ๊ตฌ์ถ ๋ฐ ๋ผ๋ฒจ๋ง ๊ด๋ฆฌ
Build and maintain custom data collection pipelines and labeling workflows for household object datasets.
๋ชจ๋ฐ์ผ ํ๊ฒฝ์ ๋ง์ถ ๋ชจ๋ธ ์ฌ์ด์ฆ INT8 ์์ํ ๋ฐ ๋ฐ์ด/๋ฉ๋ชจ๋ฆฌ ์ ์ด ์ต์ ํ
Optimize model size using INT8 quantization and improve thermal and memory efficiency for mobile environments.
์๊ณ๊ฐ ์กฐ์ ์ ํตํ AR ๋ ๋๋ง ์์ ์ฑ ํ๋ณด
Improve AR rendering stability by optimizing detection confidence thresholds and inference parameters.
Qualifications
๊ฐ์ฒด ์ธ์ ๋ชจ๋ธ(YOLO, MediaPipe, EfficientDet ๋ฑ) ์ ์ด ํ์ต ๋ฐ ํ๋ ๊ฒฝํ ์์ผ์ ๋ถ
Experience with transfer learning and fine-tuning object detection models such as YOLO, MediaPipe, EfficientDet, or similar architectures.
TFLite, ONNX ๋ฑ์ ํ์ฉํ ๋ชจ๋ฐ์ผ/์ฃ์ง ๋๋ฐ์ด์ค ๋ชจ๋ธ ๋ฐฐํฌ ๋ฐ ์ถ๋ก ๊ฒฝํ์ด ์์ผ์ ๋ถ
Hands-on experience deploying and running inference with deep learning models on mobile or edge devices using frameworks such as TensorFlow Lite (TFLite) and ONNX Runtime.
PyTorch, TensorFlow ๋ฑ ๋ฅ๋ฌ๋ ํ๋ ์์ํฌ ํ์ฉ์ ๋ฅ์ํ์ ๋ถ
Proficiency in deep learning frameworks such as PyTorch and TensorFlow.
์จ๋๋ฐ์ด์ค ํ๊ฒฝ์์์ ๋ฅ๋ฌ๋ ๋ชจ๋ธ ์ต์ ํ ์ดํด๋๊ฐ ์์ผ์ ๋ถ
Strong understanding of deep learning model optimization techniques for on-device deployment.
์นด๋ฉ๋ผ ๋น์ ๊ธฐ๋ฐ์ AR ์ ํ๋ฆฌ์ผ์ด์ ํ๋ก์ ํธ ์ฐธ์ฌ ๊ฒฝํ์ด ์์ผ์ ๋ถ
Experience developing or contributing to camera visionโbased AR applications.
์ปดํจํฐ ๋น์ ์/๋ฐ์ฌ ํ์ ์์ง์ ํน์ ๊ด๋ จ ๋ถ์ผ ์ค๋ฌด ๊ฒฝ๋ ฅ์
Master's or Ph.D. degree in Computer Vision or a related field, or equivalent industry experience.
Tech Stack
Python, PyTorch / TensorFlow, TFLite, ONNX Runtime, OpenCV
๋๊ธ