Skills
This page summarizes the capabilities most consistently reflected in public repositories, resume material, and technical output. Main Technical Areas - multimodal model related workflows - computer vision model training and evaluation - benchmark and evaluation design - data preparation and large-scale data processing - agent-related tools and engineering experiments Engineering Stack Seen Repeatedly - Python - PyTorch - TensorFlow - PaddlePaddle - OpenCV - SQL - Apache Spark - C++ and JNI in project-specific contexts Practical Strengths - turning vague research goals into structured engineering tasks - connecting data, modeling, evaluation, and deployment - building reusable tools around workflows rather than isolated demos - using repositories and notes as part of long-term technical accumulation Boundaries This page is intentionally selective. It highlights recurring and evidenced capabilities, not every framework or concept that may have been touched at some point.