Data Scientist

Thedu (தேடு) is building the business digital risk index for Asia Pacific and India — the credit-bureau model, applied to business digital risk. We measure what the outside world can already observe about a business, keep the record over time, and turn it into a score that business can act on. The platform has been running in production since April 2026 and already holds a longitudinal record of 1.6 million businesses across six markets, growing every day.Collecting the data is a solved problem here. Turning it into intelligence is the frontier — and that's this role.About the roleYou'll help build the signal layer that sits on top of the raw record: the layer that turns millions of individual observations into a business risk score people can trust. You'll work directly with the founding team and alongside the engineer who owns the data warehouse, on a dataset with a property most data scientists never get to work with — it's longitudinal. Every business is observed again and again over time, so the same record carries not just a snapshot but a history. The change over time is the signal.What you'll work onThe Business Risk Score — help design and refine the model that turns observed signals (security posture, compliance markers, infrastructure hygiene) into a single, defensible score, and validate that it actually predicts what it claims to.Longitudinal signal — build the features only a time-series makes possible: what changed, how fast, and whether it matters. This is the heart of what makes our data more than a scan.Sector and market normalisation — a café in one market and a manufacturer in another shouldn't be scored on the same absolute scale. Help classify businesses and normalise the score so it's fair across sectors and across markets.Entity resolution — match observed businesses to real-world identities so the record stays clean and the score attaches to the right entity.Turning analysis into product — your findings feed the reports and the score customers and partners actually see. The work is applied, not academic. It ships.What we're looking forWe care more about how you think than about years on a CV.A solid grounding in data science and statistics — you can frame a problem, choose a sensible method, and know the difference between a model that fits and a model that predicts.Python and the standard toolkit — pandas, scikit-learn, and comfort in SQL.Curiosity about the real world behind the data — the best signals here come from understanding what a business actually is and how its digital footprint reflects it.Rigour and honesty with findings — you name a weak result as a weak result. Here the record is the product, so intellectual honesty about the data is a core value.A builder's appetite — you want your analysis to ship inside a product, not sit in a notebook. Any exposure to time-series or longitudinal data is a genuine plus, but the mindset matters more.What you getNot the tenth data scientist tuning a mature model — an early one, on a dataset no competitor can reconstruct, with direct access to the founders and real ownership of a defined, meaningful problem. You'll be mentored by people who've spent careers in enterprise security and data, and you'll see your work go from the warehouse to a live product that businesses across the region depend on.We welcome candidates growing into a data-science career as much as those already in one.Based in Singapore. Hybrid. Check out more at thedu.io.