机器学习科学家5-预测聚合
Machine Learning Scientist 5- Forecasting Aggregation
在Netflix，我们的使命是娱乐全世界。我们正在书写下一章——突破叙事、全球粉丝和将不可能变为现实的边界。我们是一支梦想团队，痴迷于探索创意、直觉和尖端技术融合时带来的不适感与兴奋感。加入我们，成为未来的一部分。
2022年，我们推出了新的低价、支持广告的套餐，并正在打造一个内部的、世界级的广告技术生态系统，为我们的会员提供更多选择，为广告商提供优于传统电视的优质体验。我们正在寻找这个新业务领域的创始成员。
广告预测团队构建了Netflix广告业务的预测基础——在广告活动运行之前，告诉我们有多少库存可用以及广告活动将如何交付的模型。我们对受众、广告产品和格式的供需进行预测，并预测广告活动结果，如最大可用性、交付信心、覆盖范围和频率，同时考虑影响交付的广告投放优化。我们的预测支持媒体规划、承销、预算规划、收益和公共API。
这是一个全新的基础性职位。你将构建增强我们基于模拟的广告活动交付预测引擎的机器学习模型——将缓慢的基于规则的模拟转变为快速、准确、可学习的广告活动交付模型。你将全程负责建模和原型设计，并与我们的机器学习工程团队紧密合作，将模型投入生产。
### 在这个职位中，你将：
- 构建、原型设计并迭代监督式机器学习模型，用于预测广告活动交付结果——交付风险、覆盖范围、频率和竞争程度，以取代当前的模拟引擎。
- 在考虑供应方信号的同时，对需求方广告活动结果进行建模，使模型能够分析可用库存与广告商目标（定位、频率上限、竞争、节奏）之间的匹配程度。
- 设计严格的离线和在线评估框架，以衡量模型准确性、对季节性和分布变化的鲁棒性，以及相对于模拟基线的提升。
- 负责特征工程，并为团队的特征存储做出贡献——将广告投放日志、广告活动属性和供应信号转化为可重复使用、文档齐全的特征。
- 优先考虑可解释性和可审计性，确保模型决策透明且可追溯。
查看英文原文
At Netflix, our mission is to entertain the world. Together, we are writing the next episode - pushing the boundaries of storytelling, global fandom and making the unimaginable a reality. We are a dream team obsessed with the uncomfortable excitement of discovering what happens when you merge creativity, intuition and cutting-edge technology. Come be a part of what’s next.
In 2022 we launched a new lower-priced, ad-supported tier, and we are building an in-house, world-class ad-tech ecosystem to give our members more choice and to offer advertisers a premium, better-than-linear-TV experience. We are looking for the founding members of this new business area for Netflix.
The Ads Forecasting team builds the predictive foundation of the Netflix ads business — the models that tell us, before a campaign ever runs, how much inventory is available and how a campaign will deliver. We forecast supply and demand across audiences, ad products, and formats, and we predict campaign outcomes such as maximum availability, delivery confidence, reach, and frequency, accounting for the ad-serving optimizations that shape delivery. Our forecasts power media planning, underwriting, budget planning, yield, and the public API.
This is a brand-new, foundational role. You will build the machine learning models that augment our simulation-based engine for predicting campaign delivery — turning a slow, rules-based simulation into fast, accurate, learnable models of how campaigns deliver against real inventory. You'll own the modeling and prototyping end-to-end and partner closely with our ML engineering team to take models to production.
### In this role, you will:
- Build, prototype, and iterate on supervised machine learning models that predict campaign delivery outcomes — delivery risk, reach, frequency, and contention — to replace the current simulation engine.
- Model demand-side campaign outcomes while incorporating supply-side signals, so the models reason about how well available inventory matches what advertisers are trying to achieve (targeting, frequency caps, contention, pacing).
- Design rigorous offline and online evaluation frameworks to measure model accuracy, robustness to seasonality and distribution shift, and lift over the simulation baseline.
- Own feature engineering and contribute to the team's feature store — turning ad-serving logs, campaign attributes, and supply signals into reusable, well-documented features.
- Prioritize explainability and interpretability: your models' outputs must be defensible to sales and media-planning stakeholders making real booking and underwriting decisions.
- Partner with ML engineers to deploy models at scale and to monitor production model health and drift, feeding monitoring insights back into the next modeling iteration.
- Collaborate with cross-functional partners across product, engineering, and sales to define objectives, constraints, and trade-offs, and to drive adoption of ML-driven forecasts.
- Communicate technical decisions, trade-offs, and results clearly to both technical and non-technical audiences at all levels of the company.
### We are looking for:
- Advanced degree (PhD or Master's) in Statistics, Mathematics, Computer Science, or a related quantitative field.
- 5+ years of relevant experience building machine learning models on large-scale data.
- Deep expertise in supervised learning (e.g. gradient-boosted trees, regression, and related methods) with a strong bias toward interpretable, explainable models.
- Strong feature engineering skills and familiarity with feature stores and standard ML lifecycle practice (versioning, evaluation, monitoring, retraining).
- Proven ability to prototype algorithms and validate them rigorously against production data.
- Strong programming skills in Python and strong SQL.
- Working knowledge of ad-serving and campaign concepts — how campaigns are delivered and what creates delivery risk: targeting, frequency caps, contention, bidding, pacing, budget planning, and the core campaign objects/attributes; and the metrics that matter (reach, frequency, impressions, clicks, outcomes). You should understand both the supply side (ad-serving rules and inventory behavior) and the demand side (campaign attributes and advertiser goals). Ads experience is strongly preferred.
- Ability to work independently, drive your own projects, and make compelling cases for prioritization.
- Ability to communicate technical and statistical concepts clearly to audiences at many levels.
- Embodies the Netflix values while bringing a new perspective to continue improving our culture.
### Nice to have:
- Experience at a DSP, SSP, or publisher-side ad platform where predicting campaign outcomes at scale is a core science problem.
- Familiarity with our ML stack (Metaflow) or comparable large-scale ML tooling.
- Experience partnering with ML engineers to ship and monitor production ML systems.
- Experience creating data products, dashboards, or explainability tooling for non-technical stakeholders.
- Experience applying GenAI to boost developer/research productivity.
Generally, our compensation structure consists solely of an annual salary; we do not have bonuses. You choose each year how much of your compensation you want in salary versus stock options. To determine your personal top of market compensation, we rely on market indicators and consider your specific job family, background, skills, and experience to determine your compensation in the market range. The range for this role is $466,000.00 - $750,000.00. This compensation range will vary based on location.
Netflix provides comprehensive benefits including Health Plans, Mental Health support, a 401(k) Retirement Plan with employer match, Stock Option Program, Disability Programs, Health Savings and Flexible Spending Accounts, Family-forming benefits, and Life and Serious Injury Benefits. We also offer paid leave of absence programs. Full-time hourly employees accrue 35 days annually for paid time off to be used for vacation, holidays, and sick paid time off. Full-time salaried employees are immediately entitled to flexible time off. See more details about our Benefits [here](https://jobs.netflix.com/work-life-philosophy).
Netflix is a unique culture and environment. Learn more [here](https://jobs.netflix.com/culture).
[Inclusion](https://about.netflix.com/en/inclusion) is a Netflix value and we strive to host a meaningful interview experience for all candidates. If you want an accommodation/adjustment for a disability or any other reason during the hiring process, please send a request to your recruiting partner.
We are an equal-opportunity employer and celebrate diversity, recognizing that diversity builds stronger teams. We approach diversity and inclusion seriously and thoughtfully. We do not discriminate on the basis of race, religion, color, ancestry, national origin, caste, sex, sexual orientation, gender, gender identity or expression, age, disability, medical condition, pregnancy, genetic makeup, marital status, or military service.
Job is open for no less than 7 days and will be removed when the position is filled.