As a Computing and Artificial Intelligence graduate with 7 years of experience in software engineering and algorithms, I possess extensive leadership skills and a strong engineering and algorithms background. I have successfully led teams of 10-15 individuals in various areas, including implementing Chatbot systems, E-commerce Search Engines (serving billions of users), Recommended Systems, Social-Media Search Engines (with 130 million users), and AI Agent-Music-RAG (with 180 million users). With my adeptness in team leadership, robust engineering architecture building, and innovative algorithm development, I am confident in my ability to make valuable contributions to future projects.Technical Skills:- Programming Languages: Python, Java, C++, JavaScript, HTML5, SQL- Frameworks and Libraries: SpringCloud, Dubbo, Django, Flask, Protobuf, PyTorch, TensorFlow, Vue.js- AI and Machine Learning: Langchain, AI, LLM (Language Model), AIGC (Artificial Intelligence and General Computing), NLP (Natural Language Processing), Quant - Algorithm Development: Segmentor, NER (Bert-CRF), Multimodal-DSSM (Deep Semantic Similarity Model), Ranking models (ESMM - Entire Space Multi-Task Model) - Multimodal Learning: Leveraging multiple modalities for enhanced search relevance- Engineering: ES (Elasticsearch) distributed cluster, HBase, Codis cluster, Other advanced technologies- NLP (Natural Language Processing): Segmentor, NER (Named Entity Recognition), OOV (Out-of-Vocabulary) handling, SA (Sentiment Analysis), Auto-summary generation, Term-weight computation, Query corrector- Technologies: CRF (Conditional Random Fields), BiLstm-Crf (Bidirectional LSTM-CRF), Bert-Crf (BERT-CRF), Mixed-Cate-Model, CNN-FastText Model, Bert-Rnn (BERT-Recurrent Neural Network)Search Pipeline:Integration of NLP modules, recall modules, and ranking modules- Optimization: Algorithmic feature systems, Response time improvement, Efficiency enhancement- Ranking Models: 2-Tower rough sorting model, ESMM (Entire Space Multi-Task Model), PLE Models (Pairwise Learning to Rank with Multiple Pairwise Losses)- Recall Techniques: Vector Recall (utilizing Milvus), u2i (user-to-item), i2i (item-to-item), Q2t (query-to-tag), g2t (goods-to-tag)