Wei Yuan

Wei Yuan

Applied Machine Learning Scientist
Canada
English, French

About Me

Applied Machine Learning Scientist with both hands-on and research experience on NLP and GenAI, with a publication at a top-tier machine learning conference (WWW), and specialized in developing tailored models for divers…

Experience

Machine Learning Engineer (Contractor)

Branchy Solution
Jun 2023 - Present · 3 years 1 month

Developed a cutting-edge application for a recruiting agency to enhance job search efficiency by effectively matching job seekers’ resumes with available job positions, leading to an increased response rate from employers with LLM
Scheduled an ETL pipeline powered by Apy to scrape and extract data from Linkedin, Indeed, etc., and further implemented data cleaning and designed databases to enhance overall efficiency
Designed and architected a serverless computing solution on AWS utilizing Lambda and Chalice framework, further integrated Elasticsearch and Kibana for storing and visualizing the resume dataset
Implemented a resume-job matching algorithm utilizing RAG models by GPT-3.5 and Llama using LangChain, resulting in a 20% increase in response rate, while leveraging VectorDB (FAISS) for vector embedding storage and context-specific retrieval
Improved development processes by writing Pytest unit tests, connecting the CI/CD pipeline with GitHub Actions, actively monitoring model running status using AWS CloudWatch, and implementing an error-driven alarm system to ensure error detection and resolution

Machine Learning Engineer (Contractor)

Branchy Solution
Jun 2023 - Present · 3 years 1 month

• Developed a cutting-edge application for a recruiting agency to enhance job search efficiency by effectively matching job seekers’ resumes with available job positions, leading to an increased response rate from employers with LLM
• Scheduled an ETL pipeline powered by Apyfi to scrape and extract data from Linkedin, Indeed, etc., and further implemented data cleaning and designed databases to enhance overall efficiency • Designed and architected a serverless computing solution on AWS utilizing Lambda and Chalice framework, further integrated Elasticsearch and Kibana for storing and visualizing the resume dataset
• Implemented a resume-job matching algorithm utilizing RAG models by GPT-3.5 and Llama using LangChain, resulting in a 20% increase in response rate, while leveraging VectorDB (FAISS) for vector embedding storage and context-specific retrieval
• Improved development processes by writing Pytest unit tests, connecting the CI/CD pipeline with GitHub Actions, actively monitoring model running status using AWS CloudWatch, and implementing an error-driven alarm system to ensure error detection and resolution

Research Assistant

Data Mining Lab, York University
Oct 2021 - May 2023 · 1 year 7 months

Designed and implemented a novel Transformer-based Topic-Aware Response Selection Model using PyTorch, enhancing coherence and consistency in personalized dialogue system responses
Developed an innovative similarity-matching architecture incorporating word and topic-level semantics, resulting in a 1.21% on Recall@1 and 1.22% on MRR improvement by integrating a novel Topic-Aware Attentive Module
Enhanced performance on Reddit empathetic datasets happy and offmychest with improvements of 1.10% and 1.15% respectively on Recall@1 metrics

Research Developer Intern

Collaborative Intelligence Lab, Intact Financial Corporation
Jun 2021 - Oct 2021 · 4 months

Collaborated with the Customer Service team to develop bilingual conversational chatbots using cutting-edge language models in PyTorch on AWS EC2 within a Docker environment for development
Enhanced model performance by preprocessing conversational dataset using NLTK and fine-tuning Huggingface's pre-trained Sentence-BERT models over 100k conversational datasets
Improved intent detection accuracy for bilingual chatbots, leading to an accuracy improvement of 10% on English and 20% on French language models respectively

Software Developer Intern

CMEOW (Chanmao Inc.)
May 2019 - Jul 2019 · 2 months

Led the migration of product databases from PostgreSQL (on AWS) to MySQL (on Alibaba Cloud), conducting research on available services and providing the executive team with detailed options and recommendations
Designed database schema by analyzing the existing database, while following best practices for indexing and constraint for databases, and further developed Python and Bash scripts to facilitate migration

Skills

Deep Learning Natural Language Processing NLP Chatbot Language Models GenAI RAG LLM fine-tuning CNN RNN Transformer Python PyTorch TensorFlow HuggingFace NumPy Pandas Scikit-learn NLTK spaCy Gensim SQL Java Shell Script LaTeX PySpark MySQL MongoDB Elasticsearch ETL pipeline AWS EC2 S3 Lambda Chalice CloudWatch GCP Docker CI/CD Git Linux Kibana LangChain GPT-3.5 Llama FAISS Pytest GitHub Actions Bash PostgreSQL Alibaba Cloud
Report this Profile?