AI & Software Engineer - GenAI, LLMs, RAG, CV/NLP Expert

Ruochen Wang

Open to relocation. H*b visa with NIW Mobile: 302-***-**** Email: *******@*****.***

I am an AI/software engineer with 7 years of experience in artificial intelligence applications and Python development, specializing in AI agents, GenAI, RAG systems, LLM applications, Chatbot, computer vision, NLP, machine learning, and related software development. I have hands-on development experience with AI agents, workflows, data pipelines, and MCPs based on both proprietary solutions (such as Azure OpenAI, AWS Bedrock and SageMaker) and open- source frameworks (such as Llama and Qwen with vllm). Skills

Python, GenAI, LLM, Agentic AI, RAG, AI agent, CV, NLP, Langchain, OpenCV, Pytorch, AWS, Azure, Fast API, Milvus, SQL Professional Experience

Baidu – AI Engineer 2024-2025

Responsible for building a workflow-based data pipeline to produce content for AI accounts on social media like Xiaohongshu and TikTok. The workflow includes steps such as content summarization and rewriting (based on Qwen or Llama), retrieval- augmented generation (RAG using Milvus), and uses tool-calls for services like search. Delivered the module as an MCP server or HTTP endpoints for integration by other applications. Lowe’s Innovation Lab - Senior Machine Learning Engineer 2023-2024

Multimodal Chatbot: The model can utilize photos and text uploaded by users to answer specific questions and provide suggestions. The system is built on a hybrid architecture of large and small models (GPT, Llama, Yolo, Qwen-VL), orchestrated via LangGraph to manage the workflow, and uses Milvus/AWS-Bedrock with PostgreSQL to build the knowledge base.

Johnson & Johnson - Senior Data Scientist 2021-2023

Smart Reading: Engineered a highly advanced Smart Reading app that integrates cutting-edge OCR model (CRNN/ABINet) and NLP models (Bert), for seamless document recognition and analysis. Fujitsu R&D Center – AI Engineer 2016–2020

Face recognition: Work with the facial team to deliver face detection, recognition, and related services for Fujitsu's.

Document Processing System: Developed a state-of-the-art OCR system, encompassing layout analysis, text detection, and recognition. Implemented proofreading of recognition results using a language model.

Person Search: Develop a pedestrian search system in video surveillance system, capable of identifying specific individuals in crowded scenes.

Text generative model: Utilized generative adversarial networks (GAN) to generate text and characters in various styles. The generated text exhibits a more natural appearance compared to synthesized text. Education

University of Delaware 2020-2021

M.S. in Computer and Information Science

Beijing University of Technology 2009-2016

M.S. in Control Science and Engineering (Specialization in Pattern Recognition) B.S. in Electronic Science and Technology

Publication

Wang R, Zheng L, Hawke A L, et al. Video-Based 3D pose estimation for residential roofing[J]. Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2022: 1-9.

Wang R, Xia X, Zhang C, Yu X, Sun J. Text line adjustment based on neural network[C]. // 2019 15th International Conference on Document Analysis and Recognition (ICDAR 2019). Wang R, Wang S, Sun J. Offset Neural Network for Document Orientation Identification[C] //2018 13th IAPR International Workshop on Document Analysis Systems (DAS). IEEE, 2018: 269-274.

Wang R, Xu Z. A pedestrian and vehicle rapid identification model based on convolutional neural network[C] //Proceedings of the 7th International Conference on Internet Multimedia Computing and Service. ACM, 2015: 32.

Wang R, Fan W, Sun J. Method, system, and neural network for identifying direction of a document: U.S. Patent Application 15/876,334[P]. 2018-7-26.

Contact this candidate