March 23, 2021 Introducing V.1.0 of the Challenge Recommender Tool

At any given time, there are potentially hundreds of active challenges which can make it time consuming to search and browse for the ones that best fit an individual’s interest and/or strong skills. The challenge recommender is a first version of a personalized tool that is designed to suggest “Open for Registration” challenges to the members based on their past competition history. It aims to cut down on the time spent looking for relevant challenges and recommend those that give the user the highest chance at winning! 

How does it work? 

An automated Challenge Tagger Tool scans past challenges specs and automatically converts them into lists of standardized skill-related tags. If the user had challenges won in the past, the skill-tags from those challenges are combined into the user’s skill profile, and this skill profile is compared to the skill-tags of the challenges that are open for registration. A resulting match-score is used to select the contests with overlap to the user’s skills and return them in the order, sorted by the match-score, placing the most relevant contests on the top.

How do I use this tool?

You must be logged into your Topcoder account and have earned a placement in a challenge in the past 4 years.

But don’t worry if you’re new or haven’t won yet. As you gain experience on the Topcoder platform by completing skill builder challenges or placing in challenges, our system will learn and will start recommending relevant challenges that best match your skill sets.

Additional information.

Copilots will still be able to manually add tags however, currently, only the tags picked up and added by the automated Challenge Tagger are used.

What’s next?

We will enable this tool for private and enterprise communities so users can find relevant work within those sub-communities in the v2 version of the tool.

We are planning to expand and diversify the skill sources, used for Recommender, in the v3 version of the tool.

Have any feedback or questions? Go here.