Lead Machine Learning Engineer

Key Skills

Analytical and Problem solving, Data science techniques, Deep Learning, Keras software library, Leadership, Machine learning techniques, Natural Language Processing (NLP), PyTorch, Ray.io, REST API, SaaS, Spark Core, TensorFlow, Transformer

Job Description

About the team:

Klaviyo operates a real-time data analytics platform coded primarily in Python that is built for massive scale and hosted on Amazon Web Services (AWS). Engineers come to Klaviyo with experience in a variety of languages and from a number of disciplines.

About the role:

The Lead Machine Learning Engineer will help build foundational models at Klaviyo that extract insights from the massive streams of data that Klaviyo ingests continuously. You will apply cutting edge techniques from deep learning, language modeling, recommender systems, and more to generate value for our customers by allowing them to target and personalize their messages. 

The Lead MLE will be a technical leader within the organization, helping to set the standard for excellence in areas including statistical rigor, machine learning performance, and software code quality, and will amplify their own impact by mentoring other team members on technical skills.

You should have in depth experience in machine learning or data science, and have already trained and developed multiple advanced models in the recommender, NLP, or related areas. The Lead MLE must be able to lead the entire model lifecycle from inception to production. Please note that this role is based in Boston and requires a hybrid, onsite component. 

How you'll have an impact:

30 days 

• You will have finished on-boarding including engineering and data science specific on-boarding.

• Your local developer environment is set up and usable.
• You will have met multiple members of the Data Science team and are starting to familiarize yourself with both the Klaviyo product and the data sources available to you.

60 days 

• You will understand how to train and test models in Klaviyo.

• You will have developed an opinion about what’s needed to tackle your initial project including the data needed, the success metrics, the technical approach, the experiments needed to test performance in production, etc.
• You will have participated in at least 1 review for a project that a teammate is running and able to share your knowledge

90 days

• You will be actively iterating on models for your first project, and have identified opportunities and obstacles to progress with plans to address them.

• You will be actively working with teammates across Data Science to accelerate progress.
• In partnership with your manager, you will be regularly in conversation with Product and Engineering stakeholders to make sure work remains aligned with company priorities.

Up to 1 year 

• You will have developed multiple impactful models that use Klaviyo’s data, which have been put into production and are generating value in novel and powerful ways for our customers.

• You are regarded as a technical leader and have introduced new techniques to Klaviyo’s toolset that are being used by other data scientists.
• You continue to deliver new and innovative solutions to ML problems that Klaviyo faces and find new ways to generate value for our customers from Klaviyo’s data.

What we’re looking for:

• 5+ years experience in machine learning or data science
• Must have led the development of complex models on massive data sets that power customer-facing features.
• Ability to develop technical roadmaps for solving complex business problems using ML
• Experience in at least one of NLP or recommender systems
• Experience working with Transformers or other deep learning models
• Experience writing code to high software engineering standards, following best practices such as test-driven development, code review
• Experience interacting with databases, REST interfaces, and other modern SaaS development patterns
• Experience mentoring others

Nice to have:

• Personally put models into production, monitored and maintained them in production
• Experience with frameworks such as Huggingface, PyTorch, Tensorflow, Keras
• Experience with distributed training with Spark, Ray, etc.
• Experience working with customers directly
• Experience working with Product orgs