Applied Scientist III

Overview of the role:

We are looking for an Applied Scientist III to join our algorithmic and research science team. You’ll work on mathematically rigorous, research-driven problems at production scale. This role sits at the intersection of theory and application, designing algorithms that combine elegant modeling with measurable business impact. Specifically, our scientists tackle challenges across traffic shaping, fraud detection, ad quality, pricing strategies, and auction theory, along with their practical applications. We leverage the latest deep learning models alongside classical machine learning techniques to build innovative solutions.

As the heart of the InMobi Exchange, our team optimizes the company’s core business functions and creates the strategic moat that sets us apart in the market. You will not just “use models”—you will formulate them, evaluate their assumptions, tailor them to our problem domain, and bring them to life in production. Many of our challenges have no off-the-shelf solutions; we require scientific creativity to bridge research and reality.

If you thrive on solving complex, high-impact problems and want to see your ideas shape the future of a global exchange, this is the place where your work will truly make a difference.

The impact you'll make:

In this role, you’ll operate at the intersection of cutting-edge research and massive-scale production, shaping algorithms that power a global advertising marketplace, making tens of trillions of real-time decisions every day. You’ll work in an environment where models are continuously tested, evaluated, and refined — with rapid learning loops measured in hours, not weeks. Collaborating with a team that values algorithmic depth and scientific rigor, you’ll have the opportunity to prototype, publish, and deploy work that drives measurable business impact.

• Formulate, analyze, and implement algorithms that power real-time auctions, dynamic pricing, bid shaping, pacing, and traffic allocation across a massive-scale ad marketplace.
• Design and experiment with methods in online learning, reinforcement learning, multi-armed bandits, forecasting, game theory, and Bayesian modeling—in non-stationary, adversarial environments.
• Collaborate with product and engineering teams to deploy your models in production and run real-world experiments with rapid feedback loops (measured in hours, not weeks).
• Contribute to the scientific community by publishing high-quality research, conducting internal seminars, and staying abreast of advances in machine learning, algorithms, and applied statistics.
• Evaluate the long-term dynamics of deployed algorithms, incorporating feedback, exploitation-exploration trade-offs, and incentives within multi-agent systems.
• Identify new areas for innovation by translating business challenges into research questions and proposing novel, high-impact methodologies.
• Translate mathematical ideas into practical, high-performance algorithms that operate at scale in production environments.
• Explore and close the loop between model predictions and real-world outcomes, refining algorithms based on system behavior.

The experience we need:

• Ph.D. (preferred) or Master’s degree in Computer Science, Statistics, Mathematics, Operations Research, Physics, or a related quantitative discipline.
• 5.5–7 years of experience working on algorithmic or applied research problems, ideally with some production deployment experience.
• Deep grounding in one or more of:
  • Statistical learning theory, optimization, probability theory, and information theory
  • Causal inference, decision theory, game theory
  • Online learning, bandits, RL, Bayesian methods
• Strong publication record (e.g., NeurIPS, ICML, AISTATS, KDD, UAI, WSDM, EC, SODA, COLT) is a strong plus—even if not recent.
• Proficient in scientific computing with Python, including packages such as NumPy, SciPy, PyTorch, or TensorFlow.
• Comfortable working with big data platforms like Apache Spark, distributed computing, and large-scale datasets.
• A researcher’s mindset: questions first, implementation later. You are thoughtful about assumptions and rigorous about validation.
• End-to-end ownership: you can go from idea to production and thrive in applied settings.
• Prior experience in ad tech, marketplaces, or dynamic pricing is helpful but not required.