Etsy is looking for an Applied Scientist II to join the Search Retrieval team to build efficient, scalable, and intelligent retrieval systems that surface the most relevant listings from a catalog of over a hundred million unique and creative items, combining lexical matching, semantic retrieval, and hybrid models to power the first step in Etsy’s search stack.
Requirements
- You have experience developing impactful solutions to machine learning & NLP problems involving some or all of the following parts of the pipeline: data extraction, developing and training machine learning models, working the new model into production applications.
- You have good engineering and coding skills, with the ability to write high quality production code.
- Proficiency in Python, Scala, Java and/or other equivalent languages.
- You have experience working with machine learning frameworks & technologies such as Tensorflow/TFX, PyTorch, Kubeflow.
- You turn ideas into deeply reliable and well tested code that other people – or you, six months into the future – will find easy to comprehend and modify.
- Continuous and recent experience of pushing production level Machine Learning code
Responsibilities
- Designing and improving retrieval systems that power Etsy Search, incorporating techniques like dual-encoder models, vector similarity search, and query expansion
- Building large-scale indexing and serving systems for lexical and semantic retrieval
- Collaborating with applied scientists and engineers to bring retrieval models to production, optimize latency and throughput, and scale up across Etsy’s catalog
- Conducting offline and online experiments to evaluate model quality, latency, and buyer impact
- You will build and deploy scalable retrieval systems that balance relevance, recall, and performance
- You will drive innovation by bringing in state-of-the-art retrieval algorithms
- You will run various A/B experiments to improve search relevance.
Other
- stay on top of cutting-edge research in information retrieval and machine learning, and apply the best ideas in production to solve real-world challenges.
- collaborate with a cross-functional team of engineers, applied scientists, and product managers to deliver robust systems and models, and evaluate them through live A/B experiments on Etsy.com.
- You will work in a cross-functional team to prioritize projects which help achieve our product targets.
- You will analyze model performance, debug issues in retrieval behavior, and lead efforts to improve the candidate recall and diversity in search results
- You will work work closely with our product manager to help build roadmaps which range between a few weeks to 6-12 months
