Data Scientist

Why this role existsAlrightTV is a subscription-first microdrama platform. Millions of viewing events a day, across India and now Kenya — every scroll, swipe, drop off, rewatch, payment attempt and retry lands in our warehouse at event grain.It is an Instagram-shaped dataset: high frequency, short-form, behaviourally rich, and almost entirely untapped.We're looking for someone who wants to point modern AI tooling at that data and see what falls out.Not someone who writes reports. Someone who ships models that change what the product does.What you'll actually doBuild models that run in production — churn and conversion prediction, LTV, propensity-to-pay, content ranking and recommendation, watch-time optimisation. Models that make decisions, not slides.Work with LLMs and agentic workflows as a daily tool — for content understanding and tagging, for user-behaviour summarisation, for automating your own analysis loop. If you're faster because you've built your own tooling around AI, that's the point.Segment and understand a user base at scale — behavioural cohorts, engagement archetypes, early-signal detection on who converts and who churns in the first 72 hours.Own experimentation end to end — design the test, define the metric, call the result, defend it. Pricing, paywalls, onboarding, notification strategy, content merchandising.Turn ambiguous business questions into structured problems. "Why did trial conversion drop in Kenya" is the input. The output is a defensible answer and a change to the product.Automate everything you do twice. Recurring analysis should run itself.Make the numbers visible. Leadership needs to see the metrics that matter — but this is a byproduct of your work, not the job.This is not a dashboard-building role.Present to the room. You'll be in front of founders, growth, content and revenue leads regularly.Before you applyYou need to have worked with extremely large interaction datasets — billions of user events, not aggregated tables.Gaming and media/streaming backgrounds are strongly preferred, because the problems rhyme: session-level behaviour, retention curves that decay in hours, content-item interactions in the millions, and monetisation that lives or dies on the first few sessions.Social, e-commerce or consumer fintech at real scale can work too.What we're looking forIf your experience is primarily with reporting-grade or transactional data, this isn't the right fit — and that's fine, it's a specific muscle.IIT/BITS or equivalent, with 3+ years in Data Science, Product Analytics, Growth Analytics or similar.Strong Python.Comfortable moving through very large datasets without flinching at the row count.Real statistical judgement — experimentation, causal thinking, probability, knowing when a result is noise.Hands-on ML experience applied to live business problems, not coursework.Fluent with modern AI tooling and genuinely opinionated about where it helps and where it doesn't.Ability to take a vague problem, structure it, solve it, and ship the solution — with very little supervision.Communication that lands with non-technical stakeholders. Clarity over sophistication.Raw intelligence and first-principles thinking.We care far more about how you reason through an unfamiliar problem than which tools are on your CV — dashboards and BI stacks are learned in a week.What you getFull ownership of your problem space, direct access to leadership, and one of the more interesting consumer datasets in the country to work on.Small team, fast cycles, decisions made in days.