Machine Learning Engineer, Related Pins
San Francisco, CA
Pinterest helps people discover the things they love, and inspires them to go do those things in their daily lives. Related Pins, a.k.a. More Like This, is one of the largest content surfaces across Pinterest. It helps user exploration by recommending other content related to an initial pin and the user context, often helping pinners discover the content they love but can’t quite describe in words. More than 45% of pins seen anywhere on Pinterest are recommended by the Related Pins backend. See our WWW 2017 paper for further technical details.
In 2020, Related Pins takes on a new initiative to improve user session experience and optimize for downstream engagements. Our goal is to leverage cutting edge machine learning technologies to personalize the recommendation feed through diversification, click shaping, etc, to maximize the downstream impact. On this project, you will work with a group of friendly and experienced ML engineers on the full cycle of ML ranking for the Related Pins recommendation system, build the next generation ML ranking solutions to optimize the total session engagement on closeup, including bringing in personalized recommendations, diversifying user traffic, and scaling ranking infrastructure.
What you’ll do:
- Design and develop machine learning models to predict user’s intent on the closeup and enable personalized content retrieval in candidate generation.
- Provide technical vision to a group of engineers and solid IC contributions to diversify the pin-to-pin recommendations. Build necessary models and infrastructure to shape the traffic to optimize the downstream engagement, e.g. total time spent, session length.
- Partner closely with other product teams, including Shopping, Content, Trust & Safety, to experiment with different algorithms and validate their effectiveness, while gaining knowledge of how ML works in all these products.
- Use Big Data technologies (such as Hadoop, Spark, Storm) for building large scale data mining pipelines and identify product opportunities.
- Develop processes and conduct live experiments to effectively evaluate ranking quality and impact on Pinner experiences.
What we’re looking for:
- Ph.D. with 4+ years or Masters with 6+ years of software engineering/ML expertise and the ability to build scalable recommendation systems
- Hands-on experience in content retrieval, personalized ranking, and user intent prediction.
- Practical experience in machine learning, deep learning or information retrieval.
- Experience in working with cross-functional product and engineering teams to understand requirements and incorporate them in the roadmap
- Experience in working with large code bases, cross team collaboration, mentoring other engineers, giving and getting feedback, and reviewing code/systems.
- Experience with MapReduce/Hadoop and/or distributed systems.