Staff Machine Learning Engineer (Training & Inference)

About the company

ShareChat (https://sharechat.com/about) is India’s largest social media and content marketplace platform that operates exclusively in Indic languages. We empower over 300 million strong monthly active users to share their opinions, record their lives and make new friends - all within the comfort of their language of choice. We are the leaders in the social content space in India with Moj being India’s largest short-video platform (160+ million MAUs, 50 million creators), and Sharechat being India’s leading social media and content marketplace platform that operates exclusively in 15 Indic languages (180 million MAUs, 32+ million creators).

About the role

Within the Sharechat AI team, we are looking for an experienced engineer to help us further improve our recommendation system. In this role you would be responsible for developing highly distributed systems and enabling rapid ML development. You’ll join the team developing low latency pipelines making >10 billion inference requests per day, serving traffic via ML models trained on TPUs and served via GPUs.

You would be joining us at an exciting time! The science behind recommendation systems is rapidly changing, and we’re making big progress at a rapid pace.

What you’ll do

• Design and develop systems that serve recommendations to over 300 million users
• Develop scalable ML systems that enable the entire ML product lifecycle
• Improve system design and architecture to ensure high stability and performance
• Partner with peers and work in an environment that support your growth
• Collaborate with other teams to build tools and platforms to enhance ML engineering productivity

Preferred Qualifications

• Strong CS fundamentals with a track record of writing production-quality code in a modern high-level programming language (e.g. Python, Go)
• Deep expertise in automated software testing, particularly with distributed systems
• Passion for and ability to define and drive engineering initiatives to improve scalability, reliability and quality
• Strong communication skills. Experience leading projects is a plus
• Excellent debugging skills in diagnosing issues across a multi-tier architecture
• Industry experience in building end-to-end machine learning infrastructure and productionising models is a plus. Deep expertise in Golang, Kubernetes is a plus

Machine Learning at Sharechat

Serving recommendations to 300 million users entails developing large scale personalization and recommendation models that not only understand user needs and preferences, but also help 100 million+ creators grow their audiences on our platforms. A subset of the problems we tackle include:

• Create personalized feeds for 300+ million users via real-time candidate generators, multi-task prediction models, whole-page optimization, and in-session personalization.
• Nurturing our creator ecosystem, and developing models for strategic content valuation.
• Multi-objective balancing and long term measurement.

We rely extensively on state-of-the-art ML around personalization, deep learning, bandits, causal inference, optimization, ranking and recommendation.