Data Scientist - Broker Flow

Description

Data Scientist role in the Broker Flow team, based in London with hybrid working.

You will own the experiments that shape how brokers apply for finance on behalf of their clients. Brokers are one of the company's main routes to new customers.

The company provides finance for small businesses, which need funding that moves as quickly as they do. It has funded more than £4 billion for 100,000 businesses and aims to support one million SMEs.

Broker Flow sits in the third-party acquisition department. It covers a broker-referred business from before it applies to a completed credit assessment, with the aim of getting more of those businesses to an offer. A separate team, Broker Terms, sets the price, amount and duration of each offer.

The team is small and cross-functional, with commercial strategy, engineering, design and data science. It covers the full application stack, from the broker-facing API to the portal brokers use.

The team's work falls into three areas: broker engagement before an application starts, the application itself (making it easier to complete and collecting the data a credit decision needs), and operations (running the internal CRM tools and automating processes).

The company values the best idea, independent thinking, challenging untested logic and decisions based on evidence. It gives everyone access to the best AI models, which can draw on company data, and supports people in learning new tools.

Responsibilities

• Give the team evidence for its product decisions, mainly through experiment design and causal inference, with modelling and analysis alongside.
• Own each project from start to finish: frame the problem, pick the method, build the model or run the test, then present the answer to the commercial and strategy colleagues who act on it.
• Meet brokers in person to understand their context and ask better questions.
• Help shape the team's methods, since the team is new.
• Measure your impact by how many businesses reach an offer and by their lifetime value.
• Work with data scientists in the other broker and partner teams, and join the analytics guild, the company's data science community, to compare methods and build your skills.
• Share results across the business through Learnings, short write-ups presented at weekly and fortnightly sessions.
• Current project examples: refinancing (model which businesses would benefit from refinancing existing borrowing, and test a product for them); stalled applications (find where broker applications stop, and test fixes such as clearer emails, portal updates or an early offer to prompt document upload); reintroductions (test the best moment for brokers to bring existing customers back to borrow); approval probability (measure how well the estimate shown to brokers in the portal predicts actual credit decisions, and improve it); measuring impact (design early signals at sign-up and once documents arrive that predict lifetime value, so changes can be judged quickly).

Requirements

• Essential: Statistical foundations. A background in probability and statistics from a quantitative field such as mathematics, physics or statistics. Comfortable with Bayesian reasoning, and has applied hierarchical models or formal decision-making under uncertainty in real work.
• Essential: Causal inference and experimentation. You have separated signal from noise in observational and experimental data, and know where A/B testing fails and what to use instead.
• Essential: Analytical ownership. You take loosely defined problems from start to finish, framing the question, choosing the method and presenting the answer to the people who act on it.
• Essential: Data tooling. Proficient with data manipulation and modelling tools, and with scientific computing libraries.
• Essential: AI use. You have used AI to produce work you could not have delivered otherwise, such as an analysis pipeline, a prototype or a tool others use. You judge where it helps, and check its output before anyone acts on it.
• Essential: Communication. You write and speak clearly, directly and concisely, and adapt technical detail for non-technical colleagues in commercial and strategy roles.
• Bonus: Modelling depth. You have built or adapted statistical or ML models and can explain the maths behind them.
• Bonus: Stochastic methods. You have used stochastic processes for inference or time-series modelling.
• Bonus: Broker market insight. You have worked with brokers or other intermediaries in lending or financial services.
• Bonus: Coding. You write code for analysis. The team works in Python.

Conditions

Hybrid working in London.

Salary range of £70,000 to £100,000. Pay is benchmarked against market rates, with quarterly performance and salary reviews.

Flexible working hours.

Benefits

• Medical insurance through Vitality, including discounted gym membership.
• A private GP service, separate from Vitality, for you, your partner and your dependents.
• 25 days holiday per year, an extra day off on your birthday, the option to buy or sell an additional 5 days of annual leave, and unlimited unpaid leave.
• A one-month, fully paid sabbatical after 4 years.
• Immediate access to external counselling and therapy sessions for team members who need emotional or mental health support.
• 3% pension contribution on total earnings.
• An employee equity incentive scheme.
• Generous parental leave and a nursery tax benefit scheme to help save money.
• Electric car scheme and cycle to work scheme.
• Two company retreats a year, with past trips to France, Italy, Spain and further afield.
• Offices in London, Leeds, Berlin and Frankfurt, with drinks and snacks.
• Events and community groups, including running, padel, and monthly ping-pong and pool competitions.
• A learning and development budget for everyone.
• Company-wide talks with internal and external speakers.
• Access to learning platforms such as Treehouse.