Portrait of Dylan Loo

Data Engineer · Now at IBM, Singapore

I take data from raw source all the way to the decision.

Pipelines and data models, the dashboards built on them, and the presentations and stakeholder conversations that follow. Currently a data engineer at IBM, and before that supporting Google's APAC Android team, Lazada's marketing intelligence function and SingLife's sales analytics.

Worked at

Selected work

Four places I've shipped analytics.

01

Database design · ETL · Google Looker Studio · AWS · QuickSight

All data shown is dummy. The screenshots show design, not results.

CSAT & Experience reporting: data model, ETL and dashboard

I designed the database, built the ETL and designed the dashboard in Google Looker Studio for a customer-satisfaction reporting product, taking survey responses all the way to something stakeholders can act on.

  1. Survey dataPlatform survey, one cycle per half-year
  2. ETLPrepare each cycle for reporting
  3. Data modelCycle · category · sentiment · theme
  4. Looker StudioScorecard, insights, trends

Database design

  • Structured the data around the questions the dashboard has to answer: score by category, by survey cycle, against target.
  • Made category, sentiment and theme first-class dimensions, so any view can be sliced and compared without reprocessing.

ETL

  • Built the pipeline that takes each survey cycle from raw responses to reporting-ready tables.
  • Produces what the dashboard reads: category CSAT scores, the positive / neutral / negative split, and counted themes.

Dashboard design

  • Built in Google Looker Studio, leading with one number, overall CSAT against the OKR, so status reads in a second.
  • Filters for date range, category and half-year, then an insights panel written as Observation → Trend → Action.
  • Charts with the target drawn in, then ranked themes per category so every score has a visible driver.

02

via Oliver Agency

Python · PLX · ETL · Data modelling · Dashboards

Internal details removed or generalised. Numbers, takeaways and internal table names in the screenshots are blurred.

APAC Android analytics: from ad-hoc requests to a reliable data layer

I was the data analyst behind APAC Android's sales and brand reporting. The work fell into two halves: answering questions fast, and building the infrastructure so the questions got easier to answer.

70+man-hours saved per month
+5%Android retention, APAC
3×best ad-hoc speed-up vs. expected

A Analytics

  • BAU: monthly and bi-monthly update decks, with Python scripts producing the sales and branding views, and a refreshed APAC dashboard to present.
  • Core: took over a quarterly regional fact sheet. Documented it end-to-end with an ERD, then reproduced it for each subsequent quarter.
  • Strategy: led the Switch & Churn analytics project for APAC, feeding the FY25 Android retention strategy.

B Infrastructure

  • ETL & dashboarding: streamlined ETL, dashboards and reporting workflows on Google Cloud, built in PLX.
  • Data models: built and maintained models that put sales and marketing teams on the same numbers.
  • Dashboards: refreshed the APAC Android dashboard in PLX, redesigned for readability.
Screenshots shown below ↓ Deck & ERD shown below ↓
Ad-hoc requests: expected vs. actual turnaround Three requests, each from a different marketing team.

Bars share one scale, assuming an 8-hour working day. Each one was delivered with a Python script.

The data layer, simplified Redrawn generically; real table names are internal.
Sales Sales by device Sales by channel Sales by retail channel SOURCE TABLES ETL · PLX Dashboard: sales Dashboard: device DERIVED TABLES Dashboard

03

QlikSense · Snowflake · ETL · ERDs · A/B testing

Only the Cross Sell dashboard is shown, with row-level data cropped out.

QlikSense dashboards for ExCo and sales, on a Snowflake ETL I built

As a Data Analyst at Singlife, I built the QlikSense dashboards that the executive committee, the partnership distribution team and the cross-sell team worked from, and engineered the Snowflake ETL behind them.

~30% → <4%gap between OpCo and ops/finance numbers
14sheets in the Partnership Distribution dashboard
90%less manual intervention in ETL

OpCo dashboard

  • The executive committee's view of business performance had drifted from operations and finance reporting.
  • Traced how each team produced APE and VNB, found the source of the difference and fixed it.
  • Cut the gap from ~30% to under 4%, added a sheet reporting the true delta, and replaced "black box" dashboard code with a better source data model.
  • Gave senior leadership a shared view of which differences between reporting channels are acceptable.

Partnership Distribution dashboard

  • Automated the partnership distribution team's monthly reporting, which had been done by hand.
  • Built the data model linking the source data to QlikSense, and converted the existing reports into 14 sheets, from branch and adviser performance to conversion rate and top cases per BDM.
  • Saved hundreds of man-hours.

Cross Sell dashboard

  • Built during Singlife's M&A with Aviva, so the cross-sell team and ExCo could read the results of an A/B test and decide on test strategy.
  • Showed age distribution, test and control groups, campaign code and status, plus APE and VNB generated.
  • Let the team track the test through the experiment and helped ExCo pick the best model for the business.
Screenshots shown below ↓

Behind all of it: ETL on Snowflake, documented with detailed ERDs, which cut manual intervention by 90% and saved 50+ man-hours a month.

04

FBI (proprietary) · Brand tracking · Livestream · WBR

Numbers on each screenshot match the list beneath it.

Marketing intelligence dashboards for a regional e-commerce leader

As a Senior Analyst on the Marketing Intelligence team, I owned two dashboards that leadership and livestream teams used day to day, and ran the weekly business review against GMV targets.

~30%boost in key livestream metrics
+2%Consideration score growth over 1 year

Built on FBI, Lazada's proprietary BI platform. I developed and maintained both dashboards there.

Also: post-campaign analytics that shaped later marketing efforts and KPI setting.

Experience

Where I've worked.

  1. Apr 2026 – Present

    Data Engineer Singapore

    Designing the database, ETL and Google Looker Studio dashboards behind a customer-satisfaction (CSAT) reporting product.

  2. Nov 2023 – Dec 2025

    Data Analyst for Google Singapore

    • Streamlined ETL processes, dashboards and reporting workflows on Google Cloud, using PLX for ETL and dashboarding, saving 70+ man-hours per month.
    • Built and maintained data models that aligned cross-functional teams on one set of sales and marketing numbers, contributing to a 5% increase in Android retention across APAC.
    • Led the Switch & Churn analytics project for APAC, feeding the FY25 Android retention strategy.
  3. Nov 2022 – Oct 2023

    Data Analyst Singapore

    • Designed QlikSense dashboards for sales performance, agent productivity and A/B tests, increasing data accessibility by 30%.
    • Engineered ETL on Snowflake and documented it with detailed ERDs, cutting manual intervention by 90% and saving 50+ man-hours a month.
  4. Mar 2022 – Oct 2022

    Senior Analyst Singapore

    • Directed the Weekly Business Review, keeping discussions aligned to GMV targets.
    • Built the Brand Health and Livestream dashboards on FBI, Lazada's proprietary BI platform, streamlining livestream strategy for a ~30% boost in key metrics.
    • Ran post-campaign analytics that shaped KPI setting and delivered +2% Consideration growth over a year.
  5. Jun 2020 – Feb 2022

    Senior Associate, Operations Singapore

    An operations role managing e-commerce operations, where on-time delivery improved from 40% to 95%.

Education

Where I studied.

2023 – 2025

Singapore Management University

MSc, Quantitative Finance

2016 – 2020

National University of Singapore

Bachelor of Business Administration

Certifications

Certified, and still learning.

Beyond the day job

Quant curiosity and a few people-first things.

Master's project · Dec 2023

Peak/trough detection in forex

Combined signal-processing tools (Fourier transform, Kalman filter, Hurst exponent) with traditional technical indicators to pick out peaks and troughs in forex pairs.

Competition · Dec 2024

Bloomberg Trading Challenge

Finished in the top 10% globally, using sentiment analysis to trade US equities.

Leadership · NUS Kent Ridge Hall

Chairman, camp & founder of a DJ club

Led 12 people to run an orientation camp for 250 freshmen, captained a 15-player table tennis team, and founded KRemix, a DJ club I pitched and trained 10 members for.

Volunteering · 2023

Mentoring with TomoWork SG

Mentored a graduating polytechnic student with special needs.

Contact

Have a messy dataset and a deadline?

I'm happy to talk about roles in data and analytics. Email is fastest.

Download my CV

A quick check that you're human, then the PDF downloads.