Data Scientist

Murphy AI
deploys voice AI agents that help banks collect debt. We’ve built a fully-fledged platform designed to optimize recovery rates while maintaining respectful and personalized communication. Our advanced automation streamlines the process of collecting overdue invoices for banks and other industries,
providing a seamless and effective solution
.

Our AI-powered agents adapt instantly, engaging with debtors across channels like voice, messengers, email and sms to maximize results while preserving trust. By combining advanced artificial intelligence with powerful automation,
we’re setting a new standard for how businesses recover payments
.

As a fast-growing startup that has already made an impact within less than a year in the market,
we are building a talented team to scale our operations and drive our vision forward

We're looking for a Data Scientist to build the Murphy’s brain — the ML that decides who we contact, when, how and how often, to recover the most debt inside hard legal limits. Recovery is a data problem before it's a voice problem, and you own the data problem. This is one of the highest-leverage seats in the company: your models set the strategy every AI agent executes, on every call.

Murphy builds AI voice agents that collect overdue debt for banks: more recovered, faster, cheaper, fully compliant, and respectful to the debtor. We're live with Santander, Revolut, BBVA and dozens others. We've proven the product in large-scale pilots with top-tier banks — now we're making every contact smarter, and we need a scientist to own the models that decide how we collect.

• Own the contact strategy. Who Murphy contacts, when, how often, on which channel — voice, Whats App, etc. — to maximise recovery inside hard legal guardrails.

• Model the debtor. Propensity and uplift models — who picks up, who pays, who responds to which voice, script or offer — turned into live segments the agents act on.

• Optimise with experiments. A/B tests and bandits over timing, cadence, voice and negotiation policy; find what lifts recovery, and ship it fast.

• Own business-level outcomes. Collaborate with product, engineering and commercial teams to define solutions and improve the key metrics.

• You hold a degree in mathematics, statistics, machine learning, or computer science.

• Rigorous with probability and statistics. Experimental design, causal inference. You reach for the right tool, and you know when a result is real.

• AI-native, ruthless on quality. You reach for AI to automate and streamline by default — and you know exactly where it gets sloppy. You never let that through, and you control quality rigorously.

• You don’t need a collections or fintech background. We’d rather teach the domain to a killer than hire a lifer who can’t ship — you’ll pick up the domain, the data and the regulation cold, fast.

• 2+ years shipping ML that ran in production and moved business metrics. Models that pick the best next action and measure whether it actually changed the outcome; ranking, pricing or contact-strategy models. Models that ship as code, not notebooks.

• Expert Python. Comfortable in a real production codebase — you take a model from idea to production and monitor it. We use Python for ML/AI services and Type Script around it, on AWS and PostgreSQL.

• Statistics you can defend. Experimental design, causal inference, measuring the real effect of an action — not just correlations. You know when a result is real, when it’s noise, and how to test it.

• Experimentation. You’ve designed, run and interpreted A/B tests or bandits on live traffic, and shipped the winner.

• You turn models into decisions. You can explain to stakeholders why the model says to call this client at 6pm — and defend it.

Salary at the top of the benchmark.

Fair equity with enormous upside potential.

Real Ownership - Murphy's strategy is yours to drive.

Small team, high autonomy, and outsized impact.

The founding team's direct attention, and a category with a $300B+ industry behind it.

Hybrid & Flexible
:
Our default setup is hybrid – 3 days a week at our office and 2 days of remote work.

First Interview –…