Timothy Low
$ currently: data scientist @ Nanyang Polytechnic, MSBA candidate @ NUS▍
I build machine learning models, data platforms, and the analytical layer that makes them usable: from the ingestion pipeline up to the instruction layer that keeps a language model grounded.
- →ML classifiers, forecasting, leakage-safe validation
- →curated data layers and LLM grounding on Databricks
- →earlier: sales-ops reporting across 15 APAC markets at Align Technology
01 · selected work
Selected work
- 01→
Machine learning
Early-Warning Classifier for Student Outcomes
- Python
- scikit-learn
- LightGBM
- imbalanced-learn
- Jupyter
A three-class model predicting whether a student will pass, sit mid-band, or fall at risk in a module, replacing a vendor system, with leakage-safe validation and a hard focus on class imbalance.
- 02→
Data platform
Natural-Language Analytics Platform
- Databricks
- Genie
- SQL
- Python
A self-serve platform letting non-technical staff query institutional data in plain English through Databricks Genie; my work is the layer underneath: the curated tables and the model instructions that keep answers grounded.
- 03→
Data architecture
Executive Dashboard Rebuild
- Power BI
- DAX
- Power Query
- SQL
- Databricks
Inherited a bloated, unstable Power BI model behind APAC's executive reporting and rebuilt it end to end: consolidated sources, modular SQL, row-level security, and pages built around each function's needs.
- 04→
Forecasting
Quarterly Sales Forecasting
- Python
- Prophet
- SQL
- Databricks
- Power BI
A market-by-market quarterly sales forecaster for 15 APAC markets, rebuilt from an abandoned proof of concept, where most of the work was redesigning the ingestion pipeline feeding the model.
02 · what i work on
What I work on
- Machine Learning
- Classification and forecasting with leakage-safe, walk-forward validation and a hard look at class imbalance. An early-warning classifier for student outcomes; quarterly sales forecasting across 15 markets.
- Data Engineering
- The layer between raw feeds and the people asking questions: reconciled keys, curated tables, consistent revenue and FX logic, one versioned source that the model and the dashboard both read from.
- Natural-Language Analytics
- Making a plain-English query tool trustworthy: curated tables underneath, an instruction layer with enough institutional context to read questions correctly, and hard limits so it answers only from verified data.
- AI Automation
- Agentic pipelines that take on manual back-office work, turning a reference knowledge base and assessment rubrics into system instructions, with a human approval step before anything ships.
- BI & Reporting
- Executive dashboards a region runs on: consolidated data models, modular SQL, row-level security, and pages built around each function's real questions rather than a generic overview.
03 · skills & tools
Skills & tools
- Languages
- Python
- SQL
- Excel (VBA)
- ML & Modelling
- scikit-learn
- LightGBM
- Prophet
- imbalanced-learn
- Jupyter
- Platforms & BI
- Databricks
- Microsoft Fabric
- Power BI
- DAX
- Power Query
- LLM & Automation
- Genie
- Copilot Studio
- Power Automate
- AI Builder
04 · about me
About me
I'm a data scientist at Nanyang Polytechnic, where I build machine learning models and the data platforms they run on for institutional decisions.
Before this I spent three years in sales operations analytics at Align Technology, rebuilding the reporting infrastructure a 15-market region ran on. I'm also an MSc Business Analytics candidate at NUS.
$ ▍