Product · Bengaluru

You own The Ken’s data from the pipeline to the answer, and you’re the person the company trusts to say what the numbers mean, including when they say something nobody wants to hear.

Bengaluru; full time; 2 to 4 years of experience in data engineering, analytics engineering, or data science.

Link Copied

About the role

The Ken runs on judgment: which stories to chase, which products to build, which experiments to try. We want every one of those bets grounded in evidence that supports or disproves our instincts, and we have more evidence than we use. Ten years of subscribers, reads, listens, renewals, and payments sit in a warehouse, with a few hundred dbt models and dashboards on top. What we don’t have is the part that makes data trustworthy. Our jobs run on fixed times and don’t tell anyone when they fail, so a late export produces a dashboard that looks normal but is wrong. Core numbers like an active subscriber or a renewal are defined in more than one place.

We can’t say which story, newsletter, or episode brought a user who went on to subscribe, and much of what we do is meant to be judged on exactly that. We can’t run an experiment and measure it on revenue. And more of the team now asks questions of the data in plain language through AI, which makes all of this matter more: a model writing SQL against messy tables produces confident numbers that are wrong.

This role owns The Ken’s entire data, analytics, and insights stack, end to end, by themselves. Half the job is building the systems that collect, organise, and check the data. The other half is owning the answers that come out of them, and being the truth teller when an answer is inconvenient.

What you will do

Own the pipeline. The warehouse, the transformation models, and the scheduled jobs that bring in data from our site, apps, payments, and product events. Make it reliable: jobs that wait for their inputs, alerts when something fails, a daily freshness check, so nobody finds a broken pipeline by noticing a strange number.



Engineering owns how data is collected inside our products and supports you on infrastructure; everything from the moment data lands is yours. Define the numbers once. Active subscriber, signup, conversion, renewal, churn, story read, engaged day. Write each definition down, encode it with tests, and make sure every dashboard, analysis, and AI query uses the same one.

Build attribution. Which stories, newsletters, episodes, and posts bring people who later subscribe, including what they read before they signed up. It doesn’t exist today, the whole company is waiting on it, and building it is yours.

Put AI on the warehouse, safely. The team should be able to ask the data a question in plain language and get an answer that’s right. That means a clean, documented layer for AI tools like Claude to read, a metrics skill that carries the definitions, and checks that catch a plausible wrong answer before it reaches a decision.

Answer the questions that matter. Why certain stories do well and others don’t. Which product levers work. Where the funnel leaks. What predicts a renewal. Take a vague business question, break it into data questions, write the queries, and present the finding as something a team can act on.

Make experiments possible. Log who saw which variant, build a standard readout, and run the first paywall or nudge test measured on retained revenue against a holdout.

Keep the open review honest. Supply the weekly analytics series and the numbers for the monthly review. Each owner presents their own; you make sure what they present is true, and say so when it isn’t.

Protect subscriber data. Keep identified data away from the people and tools that don’t need it, pseudonymise what analysts and AI read, and keep credentials and keys where they belong.

What we are looking for

2 to 4 years working professionally with data, in data engineering,



analytics engineering, or data science, ideally at a subscription, media, or consumer internet business. You’re strong in SQL and Python, and comfortable with large datasets. Above all, we care whether you know which questions to ask, how to find the answers in data, and how to explain what they mean. You will do well here if:

– You’ve built or run a pipeline other people depended on. You can describe a time it broke and how you found out.

– You can take a vague question, turn it into data questions, and come back with a narrative. You know your way around cohort analysis, correlations, and the rest of the toolkit.

– You’ve used dbt or something like it, and you’ve written a test for a number before someone else discovered it was wrong.

– You have a point of view on LLMs and agents for analysis and automation, formed by using them, including where they fail.

– You’d rather say “this data can’t tell us that” than produce a number.

– You can tell a founder that a number they like is wrong, show why, and hold your ground.

– You think about product: how readers discover us, what their experience is, what makes them engage and share.

– Bonus: experimentation, subscription metrics (retention, renewal, lifetime value), GCP or BigQuery, data privacy.

We understand not everyone will have all of this today. If you have most of it and want to acquire the rest on the job, apply.

What you get

The whole stack and the questions that sit on top of it, at a stage when most data people own one layer of someone else’s. A warehouse and pipeline that already work, built by engineers who will support you. Ten years of subscriber behaviour at a subscription business, which is rarer data than it sounds. Founders who read data and want to be told when they’re wrong, and who see your work every month. And over a year, the experience of building a company’s measurement from definitions to decisions, with your name on the numbers everyone uses. Have questions?

Mail us at [email protected] with the name of the position in the subject line.

📌 Analytics Lead (Bengaluru)
🏢 The Ken
📍 Bengaluru