Data Scientist

Business Overview

At KINESSO, we offer a unique perspective on the marketing landscape. We're building the future of performance marketing, fueled by a dynamic data ecosystem. This includes the breadth and depth of consumer data – encompassing demographics, lifestyle, purchase behavior, and more – combined with first-party client data and the rich, real-time signals from social media platforms and DSPs. This unparalleled data fusion enables the development and deployment of sophisticated AI and machine learning models, unlocking predictive capabilities and driving highly targeted, effective campaigns. The scale and complexity of this data presents a compelling challenge for those seeking to push the boundaries of what's possible in performance marketing. The company has more than 6,000 employees operating in more than 60 countries. Learn more at www.KINESSO.com.

Position Summary

We are looking for a Data Scientist to leverage a global data ecosystem to solve complex marketing problems, using cutting-edge technologies to drive measurable impact for major brands.

The ideal candidate for this position specializes in the application of Data Science, AI and ML to problems such as campaign forecasting and optimization, real-time bidding, and video/image analysis. They have hands-on experience implementing solutions in a modern architecture stack and enjoy working in a startup-like environment, both independently and with cross-functional teams.

Key Responsibilities

Execute the Strategy:

• Collaborate with a global team of data scientists, engineers, and product managers to deliver media measurement and optimization products to hundreds of users.
• Apply Data Science, AI, ML, to campaign forecasting, cross-campaign budget optimization, and bidding algorithms in a world of increasing media fragmentation and privacy-centric measurements.
• Build data science with engineering toolkits for job orchestration, MLOps, CI/CD, RESTful APIs, and scalable infrastructure.
• Scale data science across our agencies’ client portfolio. Data science should be simple, easy-to-explain, and anchored in business value.

Innovation and Development:

• Push the boundaries of machine-assisted human decision-making, for example, by automating optimization recommendations to enhance the decision-making ability of campaign managers and media investment teams.
• Foster a culture of innovation, encouraging the exploration of new techniques and technologies such as generative AI to enhance and scale our product capabilities.

Collaboration and Communication:

• Move in lockstep with team members across Data Science, Product, and Engineering, ensuring clear and effective communication of product roadmaps, statuses, challenges, and successes.
• Collaborate closely with product teams to understand their needs and integrate AI/ML technologies into our products that meet those needs, driving value and efficiency.

Desired Skills & Experience

Key technical skills:

• Time series forecasting: Proficient in applying time series forecasting techniques, including ARIMA, Exponential Smoothing, and Prophet, using Python libraries such as statsmodels, scikit-learn, and prophet. Experienced in all stages of time series analysis, from data preprocessing and feature engineering to model validation and performance evaluation. Able to effectively communicate results and insights.
• Optimization: A strong understanding of optimization techniques is highly valued. Hands-on experience with operations research methods such as linear programming (LP) and mixed-integer programming (MIP) is a significant plus. Familiarity with optimization software packages like CPLEX, Gurobi, AMPL, or GAMS is desirable. Experience in developing custom optimization routines, potentially employing methods such as stochastic gradient descent, metaheuristics (e.g., genetic algorithms), or nonlinear programming, is also a strong asset.
• Machine Learning: Expertise in machine learning, including regression, classification, and ensemble methods like Random Forest and Gradient Boosting. Skilled in model development and feature engineering. Experience with Deep Learning is a plus.
• Statistical Analysis: Knowledge of Bayesian modelling is preferred. This may include hypothesis testing, A/B testing, synthetic control, time series forecasting, and backtesting.
• Python: Advanced knowledge of Python is essential (numpy, pandas) with the ability to write efficient, clean, and commented code for model and package development, automation processes and application development. Experience writing RESTful APIs is a plus, e.g., FastAPI.
• Data Management: Proficient in data manipulation and transformation techniques. Experience with SQL databases, ETL processes and frameworks (Dagster, Airflow, dbt), and data modeling as well as cloud-based data warehousing solutions like Snowflake or Google BigQuery. Understanding of data governance and quality control.
• MLOps: Experience in overseeing the life cycle of machine learning models from development to production. This includes observability, reproducibility and evaluation of both data and models. Experience with MLflow is a plus.
• CI/CD: Experience with agile development methodologies, such as Scrum or Kanban. Proficiency in automation tools and frameworks for continuous integration and delivery (CI/CD), and monitoring model performance in production environments.
• Cloud: Experience with cloud platforms (e.g. AWS, Azure, GCP) and containerization technologies like Docker and Kubernetes.

Experience:

• 2+ years of hands-on industry experience in designing and productionizing machine learning models and AI applications preferred (1 year industry experience for candidates with a relevant PhD).
• Master degree or higher in a CS or quantitative discipline: Computer Science/Math/Statistics/Physics/Engineering.
• Excellent communication skills, with the ability to articulate complex technical concepts to non-technical stakeholders.
• Experience with media optimization is a strong plus, especially in video or image optimization, real-time bidding and non-linear optimization algorithms. Practical experience in other industries, for example, financial time series, is welcome too.
• Demonstrated problem-solving and analytical thinking skills, with a focus on delivering practical and innovative solutions.
• Adaptability and willingness to embrace new technologies and challenges in a fast-paced and evolving environment.

IPG Mediabrands is the media and marketing solutions division of Interpublic Group (NYSE: IPG). IPG Mediabrands manages over $47 billion in marketing investment globally on behalf of its clients across its full-service agency networks UM, Initiative and Mediahub and through its award-winning specialty business units Healix, Kinesso, MAGNA, Mediabrands Content Studio, Orion Holdings, Rapport, and the IPG Media Lab. IPG Mediabrands clients include many of the world’s most recognizable and iconic brands from a broad portfolio of industry sectors including automotive, personal finance, consumer product goods (CPG), pharma, health and wellness, entertainment, financial services, energy, toys and gaming, direct to consumer and e-commerce, retail, hospitality, food and beverage, fashion and beauty. The company employs more than 18,000 diverse marketing communication professionals in more than 130 countries. Learn more at www.ipgmediabrands.com.