Data Analyst

About KKStream
KKStream is a KKBOX branch established in 2016 serving as the foremost video streaming solution provider in Asia. We provide scalable video platforms and enterprise solutions, now applied by companies including au, Japan and KKTV, Taiwan.

KKStream's culture is to highlight creativity and energetic personalities, where our people are encouraged to submit new ideas, challenge the status quo and are given the opportunity and support to achieve their full potential. We also encourage a healthy work-life balance. We are looking for a talented problem solver, who is willing to grow with us, and to help developing client-side application for our product lines.

At KKStream, decisions on how to improve product value and service business are driven by data. We are looking for a data analyst to focus on measurement, experiment design, and algorithm implementation to understand how OTT audiences engage and to improve the watching experience of millions of audiences.

You will have strong collaboration with analytic partners, domain experts, and other consumers of your work, in supporting an engineering and product function.

Role & Responsibilities

    • Drive complex analytics projects from loosely defined questions to actionable insights.
    • Design and conduct quantitative analysis for a deep understanding of how the users react to our products and platforms.
    • Constantly deliver a commentary on the result of online testing of user behavior measurement.


    • Have the technical competence to perform advanced analytics: coding skills (such as Python, Java, and Scala) and experience performing analysis with large datasets, analytics tools experience (Pandas, R, SQL, or Hadoop)
    • Have the ability to condense analysis into actionable insights to inform business decisions and influence; understand business models and high-level strategy.
    • Possess statistical competence; demonstrated experience with hands-on statistical modeling (such as regression modeling, a/b testing, significance testing etc), knowledge about machine learning (such as predictive modeling, decision trees, classification models, and clustering techniques)


    • Operating experience designing, integrating, and visualizing data report using analytic tools like Amplitude, Mixpanel, or Firebase.

Interview Process

    • Online Coding Test
    • Before you are a data scientist, you must be a programmer. There will be a coding test on codility before the interview. You will be requested to answer two questions within 2 hours. The interview will proceed depending on your result on codility. If you successfully pass, you will move on to the next stage.

    • Kaggle Assignment
    • We, KKStream, had hosted competitions on Kaggle in 2017, and made available de-identified playback data sets from KKTV. One of our Kaggle in-class competitions aimed at predicting the time when customers watched drama. Detailed description and the submission instructions are available on the competition link.
    • To complete the assignments, join the competition and get both raw and pre-processed datasets. Run the script to make predictions with pre-processed dataset. Submit the result to Kaggle and get the score. The score should be around 0.85. After that, write your own script to process the raw dataset. Your goal is to create a new pre-processed dataset just like the one we provided. After you are done, send back the source code, and MD5 of your final submission along with your Kaggle account for confirmation.

    • Presentation
    • We want to know you. Once you get the invitation to the interview, please make a slide to present things that you want to share with us. Topics you may include the following:
    • 1. Self introduction
    • 2. How you have done the Kaggle assignment
    • 3. Your pet projects, if any (Optional)
    • 4. Your experience in conducting data analysis (Optional)
    • 5. How you participated the other competitions (Optional)