Data Scientist

Pushnami is a fast-growing, Austin-based tech company powering personalized website notifications and monetization for over 10,000 sites globally. Founded in 2017, we’ve grown to deliver over 1 billion messages a day across all channels and help our partners drive substantial ROI. By joining Pushnami, you’ll be part of a dynamic team working to double our partners and visitors in 2026, while expanding into exciting new verticals and markets.

Our technology boosts performance and revenue for our partners by over 20% on average!

At Pushnami, all of our employees are seen as strategic partners who help drive company-wide outcomes. You’ll be joining an ambitious and self-motivated culture where experimentation and initiative are encouraged. We value impact over process and are looking for people who want to make a difference—not just check boxes and write code. If you're passionate about solving business problems with data and want to lead with both technical expertise and business savvy, you will be a great fit for our team!

Austin, TX is our homebase and our office is located on the North side of Austin on N Lamar Blvd next to Stiles Switch BBQ. We have breakfast and lunch catered every Tuesday, once a month on Thursday, and a fully stocked kitchen with snacks and drinks. While this position is mostly remote, we encourage in-office attendance on Tuesdays and on the Thursday preceding our monthly  3rd Friday,  when the entire company has the day off.

We're proud of the culture we've built, and our team agrees! Check out our 4.6 Glassdoor rating to hear it from them directly. But the recognition doesn't stop there:

• 6x Best Place to Work in Austin (Built In)
• #1 Fastest-Growing Company in Central Texas (Austin Business Journal)
• #2 Fastest-Growing Private Company in Texas (Inc. Magazine)
• Inc. Best in Business — Software (2022)
• Our CEO was an EY Entrepreneur of the Year Finalist (2024)

The Data Scientist will build the models and experiments that decide what we show, to whom, and what we pay for it. We route and monetize traffic across thousands of publishers — every one of those decisions is a prediction problem with revenue attached.

This role is a core member of our Data Science & Analytics team and works closely with Growth, Path, Engineering, Product, Finance, and Customer Success. The ideal candidate is comfortable owning a problem end to end: framing it with a business stakeholder, pulling the data, building the model, designing the test that proves it worked, and shipping it into production where it makes real decisions.

This is an AI-First role, meaning all of your development will be AI-assisted.

• Build models that improve performance for our advertisers and publishers. Offer and content scoring, user propensity and LTV, send-time and frequency optimization, subscriber quality prediction, churn. Our scoring service is built for segment-level iteration — you'll define segments, test strategies against them, and ship without waiting on an engineering cycle.
• Own experimentation. Design and analyze A/B tests in Growth Book, set the standards for how we call a winner, and push back when a result doesn't hold up.
• Do the causal work behind the revenue. RPV and yield analysis, sub- publisher-level performance, fraud and traffic-quality detection, incrementality on partner and channel spend. Figure out what actually drove a number instead of what correlated with it.
• Put models into production. Your work runs on AWS (Athena, Glue, Lambda, S3, DynamoDB) against real traffic, with monitoring and alerting you own. A model in a notebook isn't a result.
• Ship fast using AI tools (primarily Claude Code). You'll write prompts and AI specs, manage context, verify output, and push to production with AI.
• Partner with non-technical teams to translate their problems into data solutions and show what's possible. Much of the highest-value work here starts as a question someone in Growth or Path didn't know was answerable.

• Every engineer has either Claude Code, Codex, or Cursor Pro (company-managed licenses)
• We use AI as agents for all tasks, not just as an autocomplete tool
• W…