Team Introduction

Ginkgo recently announced a large partnership with Google Cloud to build a generative AI platform for engineering biology and for biosecurity. The AI Enablement team is responsible for delivering the ML expertise required to make this happen. We have an ambitious goal: making it as easy as possible to use AI and ML across all of Ginkgo. We have access to nearly limitless compute capacity: CPU, GPU, or TPU. We partner closely with biologists, software and DevOps engineers, and data scientists to create the necessary ML infrastructure, organize data, invent, evaluate, train, and fine-tune new models and approaches, consult and prototype new applications - whatever needs doing!

Role Description

As one of the first members of this new team, you will have a large role in molding our approaches to creating foundation models for biology, as well as creating fine-tuned and derived models for specific applications in bioengineering. You will work closely with experts in DNA design and Protein Engineering to achieve this goal.

You will be the go-to expert for deep learning model architectures and training approaches. You will often take the lead on novel model design and evaluation. Many other types of work will come your way. You may need to do data archaeology, create and debug pipelines in tools like Kubeflow or Flyte, quickly learn the basics of protein folding or codon optimization, become the company’s expert on a new tool, debug odd results created by a production model for a project under a time crunch, teach seminars and deliver tech talks, or pair with an engineer to improve the scaling characteristics of our key training pipelines. 

This is a new team, a significant company focus, and a rapidly evolving field. You will need to be able to handle ambiguity and uncertainty; on the flip side, you will be able to influence where things go and how they change.

You will identify what needs to happen, bring it to the team’s attention, and make it happen.

You will not be expected to be an expert in All The Things. You will be expected to have a high level of general technical competency, be a fast learner brimming with curiosity, and an expert in a few things - deep learning, in particular. 

Responsibilities

• Lead design and evaluation of new foundation models for biology
• Establish patterns for creating derived / fine-tuned models for specific biological applications
• Leverage Ginkgo’s extensive internally generated data, and public and in-licensed datasets of protein and DNA sequences, assay results, fermentation data time series, research archives, lab notebooks, and more. 
• Create processes for appropriate data featurization, model fine-tuning, benchmarks and evaluations, etc.
• Work collaboratively with a team of engineers and scientists
• Provide guidance, document key decisions, serve as go-to expert in Deep Learning
• Identify opportunities for application of AI and ML across the company, create prototypes, and contribute to prioritization and roadmap development for AI at Ginkgo
• Share results and learnings through means such as whitepapers, blog posts, workshops, or presentations

Minimum Requirements

• Ph.D in Artificial Intelligence, Computer Science, or equivalent (extensive practical experience may be considered in lieu of degree)
• Eight or more years of experience in industry
• Deep experience with Python.
• Expert knowledge of recent literature and state of the art for large model architectures and training approaches
• Extensive experience with building and iterating on machine/deep learning models using common frameworks such as PyTorch, Tensorflow, or JAX. Demonstrated expertise in implementing and fine-tuning neural networks, including design choices such as activation functions, loss functions, and regularization techniques.
• Excellent ability to explain complex technical concepts, in writing as well as verbally
• Experience with data preprocessing techniques, including feature engineering and data augmentation.
• Ability to collaborate effectively with cross-functional teams, including data scientists, engineers, and domain experts.
• Proficiency in using cloud computing platforms and distributed systems for training large-scale machine learning models.

Preferred Capabilities and Experience

• Familiarity with the ML ecosystem and ability to explain pros and cons of various options
• Understanding of explainability and interpretability in machine learning models, and experience with techniques like feature importance analysis or model-agnostic methods.
• Familiarity with automated machine learning (AutoML) tools and frameworks like H2O.ai or Google Cloud AutoML.
• Experience with ML and data orchestration and workflow engines like Airflow, Kubeflow, Flyte
• Experience with deploying machine learning models in production environments using containerization and deployment tools like Docker or Kubernetes.
• Familiarity with Terraform and other standard DevOps tooling
• Understanding of cloud services and platforms for machine learning and big data processing. GCP / Vertex AI experience.
• Experience operating in non-trivial cloud deployments
• Exposure to computational biology, bioinformatics, protein engineering, DNA synthesis, & more: very welcome but not required

The company maintains offices and labs in Boston, MA and Emeryville, CA. We encourage team attendance in the office on 1 team day per week, for employees within a reasonable commute distance. You will be expected to travel to Boston or Emeryville once every 6 months for a week.

Total compensation for this role is market driven, with a starting salary of $170K+, as well as company stock awards.  Base pay is ultimately determined based on a candidate’s skills, expertise, and experience.  We also offer a comprehensive benefits package including medical, dental & vision coverage, health spending accounts, voluntary benefits, leave of absence policies, Employee Assistance Program, 401(k) program with employer contribution, 8 paid holidays in addition to a full-week winter shutdown and unlimited Paid Time Off policy.