ML Infrastructure Engineer
About the job
South San Francisco, CA, USA
In person, 5 days a week
ML infrastructure engineers ensure that the data we serve to our customers is accurate, comprehensive, and relevant.
The ideal candidate has experience that spans across classical machine learning techniques (e.g. logistic regression, random forest, XGBoost), through to deep learning and large language models (LLMs). You’ve likely been following developments in LLMs closely; you have a deep interest in exploring the bounds of LLM technology, and eagerly try out the latest models as they are released. You are opinionated on how to measure LLM performance, and have built custom evaluations. You get satisfaction from exploring new datasets, and discovering new features or sources of hidden signal that help our customers understand the world and make better decisions (e.g. you probably enjoyed Moneyball).
You will:
- Own our data infrastructure for extracting structured content from a large scale document ingestion pipeline using LLMs and other technologies
- Build an evaluation framework for measuring and identifying errors in our data pipeline
- Run experiments and implement new methods to improve the performance of data retrieval and data extraction jobs
- Train predictive models on top of our structured data assets; identify new sources of valuable signal for our customers
- Work closely with customers to understand their data needs, explain our approach and address customer concerns
Qualifications:
- 4+ years of experience in industry
- Software engineer with experience in machine learning infrastructure or machine learning evaluation systems
Nice to have:
- Experience working with LLMs, particularly with LLM evals
- Experience training and fine-tuning deep learning models / LLMs
- Experience at a high growth startup
- Demonstrated interest in the life sciences; experience with applying ML to problems in biology or biotechnology
If you are excited about joining us at the earliest stage and working extremely closely with our first customers, we want to hear from you.
Email us at founders@convoke.bio.
