About Thumbtack
Thumbtack helps millions of people confidently care for their homes. Thumbtack is the one app you need to take care of and improve your home — from personalized guidance to AI tools and a best-in-class hiring experience. Every day in every county of the U.S., people turn to Thumbtack to complete urgent repairs, seasonal maintenance and bigger improvements. We help homeowners know which projects to do, when to do them and who to hire from our growing community of 300,000 local service businesses. If making an impact inspires you, join us. Imagine what we’ll build together.
About the Applied Science Team
We’re looking for applied scientists with deep expertise in machine learning, optimization, building data products, and/or statistical models. As part of a small product team you will have full ownership over your domain, so you should be a person who dreams big, then executes well.
At Thumbtack, the Applied Science team is responsible for a wide variety of problems spanning AI, machine learning, statistics, and computer science:
- Improve customer and service provider matching. Matching and optimization algorithms are fundamental to Thumbtack’s product: we now service millions of matches per week. Identifying better matches between customers and service providers has an incredible impact on the experience of customers and professionals transacting on our platform.
- Model complex relationships in the presence of many confounding factors. Predictive modeling problems are everywhere across our product. Our team works to scope, design and implement machine learning models to support Thumbtack’s product and marketing.
- Characterize marketplace dynamics. Thumbtack’s marketplaces consist of thousands of active markets across our service categories and U.S. cities. Via exploratory data analysis and experimental design, our team works to understand trends and behaviors within these markets.
- Build AI features. Use state-of-the-art approaches such as large language models (LLMs), reinforcement learning and agentic workflows to enhance customer / pro experience, increase acquisition and marketplace efficiency, and to automate and enhance internal processes.
- Build a healthy marketplace. We evolve and manage the monetization mechanics of our marketplace, including defining the parameters that affect the prices we charge.
The Challenge
Thumbtack constantly works to balance supply and demand across thousands of local markets and hundreds of service categories. The Pro Growth pod is responsible for the science behind acquiring and activating the right pros, at the right time, in the right markets, to keep the marketplace healthy and growing. As an Applied Scientist on Pro Growth, you’ll build models and experiments that shape who joins Thumbtack, how they get activated, how we forecast supply needs by market and trade, and how we accelerate time-to-value for new pros. We leverage machine learning, causal inference, and optimization to strategically grow and shape our pro supply base. Our work directly contributes to meeting demand projections, unlocking strategic partnerships, and ultimately enhancing customer experience and platform growth by ensuring a robust and high-quality supply of pros.
You’ll work shoulder-to-shoulder with engineers, product managers, and other applied scientists on a portfolio that ranges from incrementality and market forecasting to LLM-powered onboarding agents. You’ll own end-to-end: from problem framing, through experiment design, through deploying models to production, through measuring business impact.
What you’ll do
- Initiate and drive applied science projects to completion, with a focus on the business impact of those projects on pro acquisition, activation, retention, and supply health.
- Develop, deploy, and optimize machine learning models for supply-side problems, including: lead scoring and pro acquisition targeting, market-level supply forecasting, pro onboarding and activation, retention and churn prediction, and agentic onboarding workflows.
- Design and execute well-powered marketplace experiments to measure model and product impact on supply and business metrics.
- Analyze structured and unstructured data such as marketplace interactions, pro profiles, external supply signals, to identify trends, opportunities, and risks related to supply health.
- Collaborate closely with engineering, product management, business development, and marketing teams to define problems, develop end-to-end solutions, and ensure successful implementation and monitoring of ML systems in production.
- Maintain the right balance between speed of execution and scientific rigor when designing solutions for a fast-paced marketplace environment.
- Contribute to evolving the team’s agentic-first ways of working, including LLM-powered tooling, eval coverage, and the use of AI agents in day-to-day applied science.
In order to be successful, you must bring
- Master’s degree in a quantitative field (Computer Science, Machine Learning, Statistics, Operations Research, Economics, or related), or equivalent industry experience.
- 3+ years of industry experience as an applied scientist, data scientist, or ML engineer with ownership of production ML models.
- Experience in applied science within a marketplace, supply chain, or growth domain.
- Solid knowledge of machine learning techniques such as classification, regression, embedding-based approaches, and causal inference.
- Familiarity with modern LLMs (OpenAI, Anthropic Claude, Gemini, AWS Bedrock) and experience with agentic AI development practices and tools.
- Good knowledge of probability, statistics, and econometric methods, including experimental design, causal inference, and optimization.
- Experience with large-scale distributed systems (Spark, Databricks, or similar).
- Ability to effectively read, write, and debug code in programming languages such as Python and SQL.
- Ability to break down complex problems rigorously and understand the tradeoffs necessary to deliver impactful projects.
- Ability to communicate clearly and effectively to cross-functional partners of various technical levels.
- Demonstrated ability to drive seamless execution end-to-end on at least one production ML project.