Yizhe (Nick) Wen, Developer in Toronto, ON, Canada
Yizhe is available for hire
Hire Yizhe

Yizhe (Nick) Wen

Verified Expert  in Engineering

Bio

Nick is an experienced machine learning (ML) and back-end engineer. He has over three years of experience building highly scalable natural language processing (NLP) solutions. Nick is familiar with search-related technology and modern NLP implementation. He is also familiar with current machine learning operations (MLOps) and DevOps processes, and software engineering practices. Nick prides himself in translating product requirements into high-quality tech design.

Availability

Part-time

Preferred Environment

Linux, Visual Studio Code (VS Code), Python

The most amazing...

...thing I've built is an end-to-end search engine solution from the ground up.

Work Experience

Artificial Intelligence Engineer

2021 - 2022
Royal Bank of Canada
  • Architected and developed a search engine for a prospect/entity searching app. Reduced the original search latency by five times. Simplified the search logic to a unified API, using Elasticsearch, NLP, and Flask.
  • Built ETL pipelines to integrate the data from vendors. Developed directed acyclic graphs (DAG) and tailored operators to support timely data integration and monitoring. Reduced the DAG development effort, using Apache Airflow and Amazon S3 (AWS S3).
  • Developed an asynchronous RESTful APIs back end for a transaction monitoring web app in Python. Used FastAPI, SQLAlchemy, MariaDB, JSON web tokens (JWT), and PyTest.
Technologies: Python, Amazon Web Services (AWS), Machine Learning, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), Development, Elasticsearch, Apache Airflow, Big Data

Machine Learning (ML) Researcher

2020 - 2021
Messagepoint
  • Researched state-of-the-art academic papers regarding natural language generation (NLG) and prototyped multiple transformer-based seq2seq language models. Used BERT and Google publisher tag (GPT) for text summarization.
  • Designed two automatic NLG evaluation metrics. Integrated various NLP tasks (question answering, semantic similarity, constituency parsing, etc.). Reduced manual labeling cost by 90%.
  • Developed APIs for text summarization prototypes and containerized models in Docker. Deployed the prototypes on Amazon EC2 (Amazon Elastic Computer Cloud). Provided endpoints to the software team for testing.
Technologies: Amazon Web Services (AWS), Python, Elasticsearch, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), PyTorch, TensorFlow, Deep Learning

eCommerce Search Engine

Led the core ML service for query and content understanding. I provided NLP support for both real-time search requests and nightly product catalog ingestion from hundreds of retailers. The system is a complex of both ML and rule-based systems, providing swift error fixing and production fine-tuning.

Natural Language Generation (NLG): Text Summarization

Researched state-of-the-art academic papers regarding NLG). I prototyped multiple transformer-based seq2seq language models, using BERT and Google publisher tag (GPT). For the text summarization prototype model I was able to compress content length by 80% with no lost context.

Anti-money Laundering and KYC Tool for a Bank

Developed a web app that supports bank employees to easily filter, flag, and handle high net-worth clients' transactions. The ML-based program automatically performs flagging implemented to improve efficiency. The app also includes a client search engine used for the KYC process.
2019 - 2020

Master's Degree in Artificial Intelligence

Western University - London, ON, Canada

Libraries/APIs

PyTorch, TensorFlow, Scikit-learn

Tools

Apache Airflow, Celery

Languages

Python, Falcon

Storage

Elasticsearch

Frameworks

Flask

Paradigms

DevOps

Platforms

Linux, Visual Studio Code (VS Code), Amazon Web Services (AWS)

Other

Machine Learning, Natural Language Processing (NLP), Development, Big Data, Generative Pre-trained Transformers (GPT), Deep Learning, Statistics

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring