Train big models on our even bigger datasets
DescriptionThe Role:We are seeking a highly skilled and self-driven Machine Learning Engineer to join our team. In this role, you'll be instrumental in handling enormous datasets, orchestrating cloud-based computing resources, and training a multitude of advanced machine-learning models. Your work will directly contribute to our mission of creating foundational models for medicinal chemistry. While you will be dealing with massive amounts of chemical and biological information, biology and chemistry experience is not required. Our dataset can be thought of as billions of labeled sentences so experience with language models is highly relevant.
Key Responsibilities:- Manage and optimize data processing workflows for large-scale datasets, with an approach akin to language data handling.
- Scale and maintain machine learning model training processes, with a focus on cloud environments (primarily Google Cloud, with flexibility to other platforms).
- Collaborate closely with ML researchers, data scientists, and lab automation teams to ensure seamless integration of lab data and ML model training.
- Innovate and iterate on our existing technology stack, taking the initiative to solve problems and improve our ML operations.
- Act as a self-sufficient project manager, overseeing your projects from conception to completion.
About You:- Strong experience in machine learning engineering, including data handling, model training, and scaling in cloud environments.
- Comfortable building ML infrastructure
- Experience working with large amounts of text data, NLP, or training LLMs
- Demonstrated capability to make informed decisions, take ownership of solutions, and drive projects forward in a startup environment.
- Excellent collaboration skills, with the ability to work effectively with cross-functional teams.
Preferred Qualifications:- Familiarity with common MLops tooling (e.g., Dagster, Prefect, Airflow, Docker, MLflow, Kubeflow, W&B, Ray, etc.)
- Ability to manage own compute cluster
- Ability to maximize GPU utilization and keep cluster busy 24/7
- Ability to analyze model results and kick off new experiments in response
- Experience with BERT or similar language models in PyTorch.
- Experience or interest in biology, chemistry, or related fields is a plus.
Salary$150,000 - $200,000 per year
