Experience

Karhoo

• United Kingdom
• IT Services and IT Consulting
• 1 - 100 Employee

• Senior Data Engineer

  • Oct 2021 - Present

  London, England, United Kingdom - Initiate and lead the design and implementation of MLOps system using tools such as Kubeflow Pipelines and MLflow tracking server. The system allows multiple models to be trained simultaneously in a Kubernetes cluster and have their metadata and metrics such as accuracy tracked and stored in the MLFlow server. Github integrations were also adopted to allow automatic model deployment. - Collaborate with Data Scientists to design and create a Python application to allow end users… Show more - Initiate and lead the design and implementation of MLOps system using tools such as Kubeflow Pipelines and MLflow tracking server. The system allows multiple models to be trained simultaneously in a Kubernetes cluster and have their metadata and metrics such as accuracy tracked and stored in the MLFlow server. Github integrations were also adopted to allow automatic model deployment. - Collaborate with Data Scientists to design and create a Python application to allow end users triggering ML model training jobs directly. The result is a significant decrease in turn around time for ML model training and deployment (1 or 2 days instead of weeks) - Adopted the idea of data observability to ensure the schema, volume and distribution of data are as expected post ETL. This increases the reliability of data for other downstream purposes such as ML model training or BI analytics reporting

• Data Engineer

  • Aug 2019 - Oct 2021

  London, England, United Kingdom - Schedule and configure DAGs in Apache Airflow to perform ETL jobs with various file formats, destinations and communication protocols. For example querying data from the data warehouse and storing transformed results as CSVs in a remote SFTP server. In the same time ensuring jobs data operations adhere to GDPR standard - Create and design helm charts and docker containers to server ML models via the Seldon framework. The result provides useful features in the deployment such as out of… Show more - Schedule and configure DAGs in Apache Airflow to perform ETL jobs with various file formats, destinations and communication protocols. For example querying data from the data warehouse and storing transformed results as CSVs in a remote SFTP server. In the same time ensuring jobs data operations adhere to GDPR standard - Create and design helm charts and docker containers to server ML models via the Seldon framework. The result provides useful features in the deployment such as out of the box model monitoring and flexible request handling such as routing to shadow models. - Work in a team to design, create and maintain other Python applications and APIs to assist in business development. For example, a complex application that could demonstrate the platform’s capability to potential business partners



Focaldata

• United Kingdom
• Market Research
• 1 - 100 Employee

• Data Engineer

  • Jul 2018 - Aug 2019

  Designing and building machine learning pipeline with Apache Airflow and Kubernetes that supports over 200 machine learning model running simultaneously. Adopting the practise of "infrastructure as code" which currently covers around 70% of the cloud resources. Enhancing infrastructure security by deploying OAuth2 authentication solutions and search engine indexing guidance.

• Junior Data Scientist

  • Apr 2017 - Jul 2018

  London, United Kingdom Deployed components in machine learning pipeline that enabled auto variables and levels matching. This reduced human error and accelerated project flow by an average of 1 day. Implemented complementing unit and integration tests for CI/CD system to ensure solutions were robust.



General Assembly

• United States
• Higher Education
• 700 & Above Employee

• Data Science Immersive Student (DSI 1)

  • Oct 2016 - Jan 2017

  London, United Kingdom The course was taught with combination of lectures and hands on practises including presentations. Assignments were designed to develop various practical skills of the student. They range from web scraping jobs information from the web to the analysis of a popular game PokemonGo! During the 12 weeks immersive course I gained a good experience with the following technologies: - Data gathering (Web scraping, API accessing, risk and assumption of data) - Data cleaning (Jupyter… Show more The course was taught with combination of lectures and hands on practises including presentations. Assignments were designed to develop various practical skills of the student. They range from web scraping jobs information from the web to the analysis of a popular game PokemonGo! During the 12 weeks immersive course I gained a good experience with the following technologies: - Data gathering (Web scraping, API accessing, risk and assumption of data) - Data cleaning (Jupyter Notebook, Python, Pandas, Numpy, Scipy) - Machine Learning (Scikit Learn, Statsmodel) - Results presentation and communication (Tableau, Matplotlib, Seaborn) Show less



HAECO

• Hong Kong
• Airlines and Aviation
• 700 & Above Employee

• Engineering Trainee

  • Sep 2013 - Feb 2016

  80 South Perimeter Road Hong Kong International Airport Lantau, Hong Kong HAECO engineering trainee programme was a comprehensive experience that allows graduates to demonstrate their capabilities through various projects in different departments . From hands on aircraft repair in difficult situations (3 a.m in the morning in the winter) to on-boarding new members regarding planning and production-control workflow, these experience played an important role in both my personal and career development.



PARSONS BRINCKERHOFF

• Summer Intern

  • Jul 2011 - Sep 2011

  7/F One Kowloon, 1 Wang Yuen Street, Kowloon Bay, Kowloon, Hong Kong • Worked in a team of 8 to improve on the process in calculating the air conditioning load of a building which resulted in a quicker analysis process (around 15% time saved using VB macro in MS Excel) • Analysed the prospects of government’s data center project.



Precise Computer Tooling Co. Ltd

• Summer Assistant

  • Jul 2009 - Aug 2009

  Huizhou, Guangdong, China • Conducted research in alternative fire redundant material for cost lowering. • Joined quality control team in product inspection • Discussed with director about improving efficiency of the manufacturing process