Senior Machine Learning Engineer

The Role

At Depop, machine learning is integral to our value proposition. In the ranking team, we build learning-to-rank models that power personalised experiences across the Depop app (in search results, recommendations, etc.).

The team is currently made of 3 Machine Learning Scientists and 1 Machine Learning Engineer. It owns a series of models deployed in SageMaker for real-time inference. These models are called in by various services across the app (e.g. the search service to rank results coming from our vector database in OpenSearch), serving millions of personalised results to users daily.

We are looking for a dedicated Senior Machine Learning engineer to join our Ranking team. As part of this team, you will participate in building, deploying and monitoring the future ranking models that will improve user experience across the app.

Responsibilities

1. Design and implement pipelines for training, deploying & monitoring real-time ranking models, in collaboration with the other ML Engineer(s) in the team.
2. Work closely with ML Scientists in the ranking team on the experimentation and deployment of new models.
3. Collaborate with Backend Engineers from "client services" (e.g. search service, which calls one of the real-time models) to define requirements and plan future experiments.
4. Help design and build the ML platform at Depop in collaboration with the MLOps infrastructure team, working on various areas:
1. Robust prototyping & training of models
2. CI/CD pipelines for model deployments
3. Model serving for real-time and batch implementations
4. Improving our feature store to serve features offline/online
5. Monitoring & alerting
5. Hold high standards for operational excellence; from running your own services to testing, monitoring, maintenance and reacting to production issues.
6. Contribute to a strong engineering culture in the ML group, orientated on technical innovation, and professional development.

Requirements

1. Consistent track record of building pipelines to train & deploy ML models and contributing to an ML platform
2. Experience with the core concepts of data science / ML workflows
3. A strong sense of ownership, autonomy and a highly organised nature.
4. Outstanding communication skills, especially in taking care of multiple stakeholders
5. Solid understanding of systems design within a modern cloud-based environment (AWS, GCP)

Technologies and Tools

1. Python
2. Data science / ML / MLOps tooling: e.g. Sagemaker, Databricks, TFServing and more
3. Common ML libraries: scikit-learn, pytorch/tensorflow, mlflow etc.
4. Spark & DataBricks
5. AWS - IAM, S3, redis, ECS and more
6. Shell scripting and related tooling
7. Good working understanding of continuous integration/deployment tools and practices
8. Experience with streaming and/or batch-based systems supporting data integrations to third-party platforms (e.g. using Kafka, Airflow, RMQ, etc.)