Its technology lets advertisers reach the right people across the open internet, including streaming TV, podcasts, mobile apps and news sites.
Ads fund much of the content people read and watch. The company aims to make advertising more transparent, effective and responsible, which supports journalism, entertainment and creators worldwide. Brands and agencies use the platform to reach customers and grow.
The platform's scale creates hard technical problems, such as processing huge datasets in real time and running reliably worldwide. Your work has a global reach. The team values varied viewpoints, curiosity and learning from each other.
The role: The pacing system decides in real time how to spend each campaign's budget across its flight, handling 20M queries per second. It must bid the right amount on each impression at the right rate, so advertisers spend the full budget and get good value.
Standard A/B traffic splits fail here, because both arms draw on the same campaign budget and one arm's spend affects the other. The team uses budget split experiments instead. Each arm gets its own budget share, pacing calculation and spend tracking, all within one live campaign. This is now the main way the team checks any change to bidding behavior.
You will be a senior engineer covering two areas: the pacing engine and the experimentation infrastructure used to test changes to it. You will work on the services that compute pacing instructions and apply them at bid time, and extend the experiment framework so any team can launch a sound experiment in hours rather than weeks.
The team works closely together and uses a wide stack, including.NET, Kafka, Kubernetes, Spark and Delta Lake. Experience with experimentation platforms or real-time decisioning outside ad tech, such as trading, payments or marketplace pricing, is welcome.
Responsibilities
Own pacing features from design to production, covering the services that set budget allocation and pacing limits on a continuous cycle and the bid-time code that applies them.
Extend the budget split experiment framework: deterministic user assignment, per-arm budget partitioning and enforcement, spend attribution and the full experiment lifecycle.
Expand budget split experiments to more of the bidding stack and campaign types, and shorten launch time.
Build guardrails that keep experiments safe on live customer budgets.
Build data pipelines, metrics and dashboards so engineers, data scientists and product managers can compare arms on delivery, pacing and campaign KPIs and make ship decisions.
Work with applied science and product on pacing and bidding algorithm changes, and validate them through A/A and A/B experiments.
Lead complex changes across several components and teams, including projects with high technical ambiguity.
Ship quality changes with automated tests, metrics, monitoring and staged rollouts.
Understand the business goals behind the code and help define clear success criteria.
Use data to analyze and explain the effect of your work.
Mentor others, set a standard through code review and help with hiring.
Requirements
Minimum qualifications: 5+ years of software development experience.
Proficiency in at least one of C, Java, C++ or Python, plus working knowledge of SQL.
Experience building and running distributed systems at scale.
Solid grounding in computer science fundamentals.
Experience with data processing and querying large datasets to debug and check system behavior.
A record of owning software from design through production, including testing, monitoring and on-call support.
Clear communication with engineering, product and data science partners.
Preferred skills
Experience building or running an experimentation platform, covering assignment and bucketing, exposure logging, guardrail metrics and analysis pipelines.
Working knowledge of experiment design and statistics, such as A/A tests, power, variance and interference between arms, especially in marketplaces or budget-constrained systems.
Experience with real-time decisioning or control systems such as pacing, bidding, rate limiting, trading, payments or pricing.
Experience with low-latency, high-throughput services.
Stream and batch processing with Kafka, Spark, Databricks or Delta Lake.
Experience with databases such as Microsoft SQL Server, Vertica, Aerospike or Redis.
Experience with Kubernetes and cloud infrastructure.
Experience with observability tools such as Grafana and Prometheus.
Ad tech experience is helpful but not required.
Conditions
Base salary range: $124,900 - $228,900 USD. This is the company's reasonable estimate of base pay under various US state laws. Actual pay may vary based on non-discriminatory factors such as experience, knowledge, skills and location. Base salary is one part of total compensation.
Sick time accrues at 1 hour per 30 hours worked.
Full-time employees get vacation of up to 120 hours in the first year and 160 hours after that, plus around 13 paid holidays per year.
Interns are not eligible for variable incentive awards, stock-based compensation, retirement plan, vacation, tuition reimbursement or parental leave.
The company is an Equal Opportunity Employer committed to an inclusive hiring process.
Benefits
Comprehensive medical, dental and vision coverage, with premiums paid in full for employees and dependents.
Retirement benefits, including a 401k plan with company match.
Short- and long-term disability coverage and basic life insurance.
Well-being benefits and reimbursement for certain tuition expenses.
Parental leave.
Eligibility for stock-based compensation grants, based on company and individual performance.
Variable compensation, incentives and commissions depending on the role.
Eligibility to buy company stock at a discount through the Employee Stock Purchase Plan.