Senior Applied Scientist-Measurement
About The Position
The Trade Desk is a global technology company and the world’s leading independent platform for digital advertising, with nearly 4,000 employees across more than 30 offices. Our technology helps advertisers reach the right audiences across the open internet — from streaming TV and podcasts to mobile apps, news, and more. Advertising powers the content people love. By making it more transparent, effective, and responsible, we help support trusted journalism, quality entertainment, and creators worldwide. The world’s brands and agencies rely on us to reach their customers and grow their businesses responsibly. The scale of our platform brings unique technical challenges — from processing massive datasets in real time to building systems that operate reliably on a global scale. When you work here, your impact is worldwide. We welcome diverse perspectives, encourage curiosity, and build teams that learn from one another. If you’re driven to solve meaningful challenges, we’d love to meet you. Applied scientists at TTD work closely with engineering throughout the lifecycle of the product, from ideation to production and monitoring. Our applied scientists are end-to-end owners. You will participate actively in all aspects of designing, researching, building, and delivering data-focused products for our clients and traders. Bringing objective, transparent measurement to the open internet is core to TTD’s strategy, and retail conversion data is an essential part of that. This role owns that data. This particular role is responsible for the research and application of state-of-the-art statistical and modeling techniques to solve measurement problems centered around retail conversion data. This role will be the team’s expert on retail data — understanding the ins and outs of every source we use, how each is collected and where it can mislead, and owning the statistical integrity of these datasets end to end. Day to day, this role will explore new data sources, build new-buyer and conversion reports, project and scale attributed numbers up and down defensibly, impute missing data, monitor data quality, and reason carefully about how retail data behaves inside various measurement models. The work of this role ensures our retail measurement solutions are statistically sound and robust, and it helps steer product decisions toward methods that hold up.
Requirements
- Proficient in Python, SQL, and PySpark, with a strong passion for enhancing and expanding your technical skills.
- Strong at data processing — exploring, cleaning, and transforming large, messy datasets — and have a deep understanding of the foundations of statistics, including estimation, sampling, and measurement error.
- Hands-on experience building statistical solutions and data pipelines at scale, with a track record of owning a project end-to-end (from research to production) and partnering with a cross-functional team of scientists, engineers, and product managers.
- Comfortable becoming the go-to expert on a complex, messy dataset.
- A keen sense of data intuition and statistical rigor: you can reason about how a data source biases a result, defend a number that cannot be directly measured with an honest error bound, and tell a solution that is statistically sound and robust from one that only looks good in the short term.
- BS/MS with 4+ years or a PhD with 2+ years of experience working in a DS or ML role that involves bringing products from ideation to production.
- Experience working with retail, panel, survey, or conversion data, and reasoning about how a data source biases a downstream estimate (match-rate composition, coverage and panel skew, deduplication).
- Experience estimating population quantities from partial or biased samples: projecting an observed or matched subset up to a full population via sample weighting or calibration, and attaching honest error bounds (e.g. via resampling).
- Rigorous missing-data imputation practice that carries the added uncertainty through to the final estimate, with the judgment to recognize when data is missing in a way no imputation can fix.
- Proficient in Python and SQL.
- The ability to communicate with diverse stakeholders, making architecture recommendations, ensuring effective execution, and measuring quality of outcomes
Nice To Haves
- Achievements like first-author publications or clear project successes are a plus.
- Experience building monitoring, anomaly detection, and data-quality checks on production metrics and data feeds is a plus.
- Experience in causal inference and lift measurement is a plus.
- Experience in programmatic advertising is a plus.
- Experience running heavy workloads on a distributed computing cluster (especially EMR or Databricks), leveraging technologies like Spark to work with large datasets preferred.
Responsibilities
- Own our retail conversion datasets end to end — be the team’s expert on these sources, how they are collected, and where they can mislead.
- Explore and process data from a variety of retail sources, building the datasets and pipelines that downstream measurement relies on.
- Apply rigorous statistics to scale and project our measurement numbers, impute missing data, and produce new reports — always with careful attention to bias and uncertainty.
- Monitor data quality and the stability of our metrics, distinguishing real shifts from noise.
- Reason about how retail data feeds conversion lift, geo lift, and attribution, and validate that our methods are robust on real data.
- Partner with cross-functional stakeholders and communicate learnings in compelling ways that steer product toward statistically sound, robust solutions.
Benefits
- Comprehensive healthcare (medical, dental, and vision) with premiums paid in full for employees and dependents
- Retirement benefits such as a 401k plan and company match
- Short and long-term disability coverage
- Basic life insurance
- Well-being benefits
- Reimbursement for certain tuition expenses
- Parental leave
- Sick time of 1 hour per 30 hours worked
- Vacation time for full-time employees up to 120 hours thru the first year and 160 hours thereafter
- Around 13 paid holidays per year
- Employees can also purchase The Trade Desk stock at a discount through The Trade Desk’s Employee Stock Purchase Plan.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
