Data Scientist (Mediamath)
MediaMath is a technology platform that brings together all forms of digital media, massive amounts of data, and sophisticated algorithms to power smarter marketing for the world’s leading advertisers. Technology like real-time bidding is revolutionizing marketing and MediaMath is at the cutting edge. Our platform handles millions of transactions per second and we reach hundreds of millions of desktop and mobile internet users worldwide. Our engineers develop complex, innovative, and highly scalable technology that is changing the face of digital media – devising new approaches, inventing new products, and opening up new markets. It’s a very exciting company in a very exciting industry. We’ve enjoyed massive growth since our founding in 2007, and we’re now a global company with offices in 20 locations – but this revolution has just begun!
We are currently looking for a Data Scientist to support the ongoing development of MediaMath’s proprietary algorithms and analytics. This individual will be a member of the Data Science team, working closely with the Product & Engineering teams on the conception, design, development, testing, and deployment of real-world applications of models & data that impact billions of dollars of marketing spend. From optimizing real-time bidding auctions, to separating human from non-human web traffic, to building out a global cross-device graph across billions of users, this individual will have the opportunity to work on numerous cutting-edge problems and develop scalable, high-performance solutions to big data problems using state-of-the art technologies, languages and frameworks.
- Develop a solid understanding of MediaMath’s business context, key problems and issues of importance to the company, and the ways in which data-driven models and algorithms can help drive performance and provide insights.
- Develop a deep understanding of business data-sets through a combination of database queries, and exploratory statistical analysis and visualization.
- Rapidly develop a deep understanding of the quantitative methodologies and implementation details behind existing machine learning algorithms in place, and the currently-planned roadmap of enhancements.
- Formulate business needs as data science problems where applicable and effectively communicate these to peers, managers, and key stakeholders, including the benefits and limitations of models.
- Design and develop Machine Learning models and algorithms that drive performance and provide insights, from prototyping to production deployment, across key areas of interest to the company (e.g., bidding optimization, messaging optimization, multi-touch attribution, fraud detection, device identification, cross-device association)
- Partner closely with Engineering on the architecture and implementation of modeling efforts to ensure performance and scalability
- Partner closely with Product on the incorporation of new modeling features into our product set, including UI & API layers
- Develop tools and processes to monitor performance of existing models and implement enhancements to improve scalability, reliability, and performance.
- Create supporting documentation for algorithms and models.
- Stay abreast of new developments in Machine Learning and Data Science, and investigate & develop new “challenger” approaches to compete against current “champions”
- Communicate effectively with other Data Scientists, with Product/Engineering, and with business stakeholders across the company to identify and define strategic data-intensive projects based on business needs.
- Take ownership of significant large projects, develop an overall architecture and roadmap, and supervise junior Data Scientists or Engineers.
- Act as a subject matter expert on areas of ownership, for both internal and external audiences
- Contribute to MediaMath Data Science thought leadership via publication of new-to-the-world research in relevant journals, blog entries, etc.
- BS, Masters, or PhD degree in a quantitative discipline (e.g., Computer Science, Math, Physics, Statistics, Electrical Engineering, or similar)
- Strong quantitative skills, with solid grasp of key concepts in Probability, Statistics, Algorithm design, and Machine Learning.
- Strong communication skills, and the ability to effectively discuss models with other data-scientists as well as business partners at the appropriate level of technical detail.
- 3+ years of hands-on experience developing software, with at least a year in Scala and/or Python.
- 3+ years of hands-on experience with machine learning, predictive modeling, and statistical analysis techniques in a business environment.
- Passion for hands-on “in-the-trenches” work with real-world data-sets.
- Desirable: Experience with Machine-Learning/Big-Data Platforms and modeling frameworks, especially Spark and Hadoop.
- Desirable: Experience with Statistical modeling and Visualization with R or Python
- Fast learner
- Analytic thinker
- Creative problem solver
- Passion for applying quantitative methods to rea-world problems
- Highly motivated
- Collaborative team player
- Focus on impact and scalability
- Direct, honest communication style
MediaMath is a global technology company that’s leading the movement to revolutionize traditional marketing and empowering marketers to unleash the power of goal-based marketing at scale, transparently across the enterprise. Our platform – TerminalOne Marketing Operating System – handles billions of transactions every hour and hundreds of millions of internet users every day, which means every solution must be built to scale. Our breakthroughs create new marketplaces and solve long-standing problems in an industry that is constantly evolving. Our engineers are building the leading technology platform to power the new digital marketing ecosystem, and we are looking for driven, curious innovators to join our team.
In achieving their duties and responsibilities, MediaMath employees embody the Math Values of SPACE: Scalable Innovation, Performance, Accountability, Collaboration, and Empowerment.
Job posted 2/11/2016