Dr. Vincent Li

Dr. Vincent Li

Data and AI/ML Lead

Vinceli.org

Programa

Biography

Dr. Vincent Li is an applied machine learning and search scientist with a focus on building AI systems that solve real-world problems. He is currently the Head of Data and AI/ML at Programa, where he leads a multidisciplinary team across data engineering, analytics, and machine learning to power smart tools for interior designers and suppliers.

Previously, Vincent held impactful roles at Canva, SEEK, and Coles, where he worked on large-scale search, recommendation, and analytics systems. At Coles, he led the redevelopment of the search engine for Australia’s largest supermarket using modern AI and ranking techniques. His experience spans the entire machine learning lifecycle—from data platform and model deployment to experimentation and impact measurement.

Vincent holds a PhD in Computer Science from RMIT University, where he researched AI methods for ranking and search. He is the organizer of the Melbourne Search & Recommendation Meetup and frequently speaks at conferences and universities about applied AI, MLOps, and product-focused machine learning.

Interests

  • Artificial Intelligence
  • Computational Linguistics
  • Information Retrieval
  • Recommender Systems
  • Ranking Problems

Education

  • PhD in Computer Science, 2016

    RMIT University

  • Master with Honor in Computing, 2014

    Australian National University

  • BSc in Software Engineering, 2007

    Soochow University

Skills

R

70%

Python

70%

Statistics

80%

Machine Learning

60%

Search

60%

Photography

20%

Experience

 
 
 
 
 

Head of Data and AI/ML

Programa

Jan 2024 – Present Melbourne

Leading the data and AI/ML function to build the infrastructure, analytics, and machine learning systems behind the world’s largest interior design product database. Responsibilities include:

  • Building AI-powered product discovery and search systems
  • Leading a team across data engineering, analytics, and ML
  • Driving AI vision and strategy across multiple product verticals
 
 
 
 
 

Founder & Principal Consultant

Aperios Lab

Jan 2023 – Dec 2023 Melbourne

Ran a boutique consulting practice focused on AI and search. Responsibilities included:

  • Helping startups and enterprises design and evaluate search and recommendation systems
  • Providing strategic and technical guidance on data platforms, MLOps, and productized ML
  • Delivering end-to-end solutions and workshops on applied AI topics
 
 
 
 
 

Senior Data Scientist

Canva

Mar 2022 – Jan 2023 Sydney

Responsibilities include:

  • Online and offline evaluation for search and recommendation systems
  • Applying state-of-the-art machine learning techniques to search relevance
  • Enhancing the data platform for search and recommendation tracking and analysis
 
 
 
 
 

Search Science Consultant

Coles Ltd

Aug 2020 – Feb 2022 Melbourne

Responsibilities include:

  • Building AI-powered search engines for retail applications
  • Applying machine learning and AI to improve search, matching, and user satisfaction
  • Designing and running scientific search evaluation frameworks
 
 
 
 
 

Applied Search Scientist

SEEK Ltd

Oct 2017 – Apr 2020 Melbourne

Responsibilities include:

  • Search engine performance evaluation and optimization
  • Conducting research to understand user behavior and improve search systems
  • Designing metrics and experiments for robust search evaluation
 
 
 
 
 

Researcher

Abbrev8

Oct 2016 – Jan 2017 Melbourne
Conducted research to evaluate and improve salient entity recognition algorithms for the company’s NLP platform.

Recent Posts

Multi-stage ranking for relevance and user satisfaction for search

Ranking is a fundamental and critical component of modern search engines. The effectiveness of ranking is directly linked to the search relevancy and user satisfaction.

Why search is so hard

Search is one of the key areas where big companies have spent lots of investment on. Working as an applied search scientist, I was always questioned why we still spend so much time on a problem that has been solved by many platforms like Elasticsearch, Luence and many others? The truth is, unless you are familar with the area, you probably won’t realise how challenging the problem is.

Recent Publications

Quickly discover relevant content by filtering publications.

Self-labeling methods for unsupervised transfer ranking

Merging algorithms for enterprise search

Contact