Machine Learning Engineer

As a machine learning engineer at NE47 Bio, you will be responsible for designing, implementing, training, and deploying machine learning models for our OpenProtein.AI platform. As part of the machine learning team, you will analyze protein sequence, structure, and function data to develop new algorithms, benchmark and evaluate existing algorithms, and deploy those algorithms in the cloud. You will work closely with the backend and frontend teams to ensure that models are productionized and deployed efficiently, and that users are able to interact with and easily interpret model predictions. You will also work closely with management and other stakeholders to understand scientific requirements and plan and develop new features and the algorithms behind them. As part of this, you should expect to work with state-of-the-art machine learning models (Transformers and RNNs), Bayesian methods (Gaussian processes, ensembles, and Markov-chain Monte Carlo), and optimization techniques for sequence design (genetic algorithms and Bayesian optimization).

Your responsibilities will include:

- Implementing, training, debugging, optimizing, and deploying large scale machine learning models
- Researching and implementing systems for efficient deployment of large models into production (e.g. compression, distillation, quantization) and integrating them with other learning algorithms (e.g., fine-tuning, transfer learning)
- Researching and implementing systems for active learning, experimental design, and discrete optimization
- Learning about and implementing state-of-the-art methods for protein sequence and structure analysis
- Writing clean, extensible code
- Data preprocessing and curation to develop training datasets and benchmarks
- Working closely with front
- and back-end teams and other stakeholders

**Job Requirements**
- Bachelor’s in Computer Science, Statistics, Bioinformatics/Computational Biology, or related fields. Masters or PhD is preferred.
- At least 5 years of machine learning experience or higher degree
- Experience with Python and deep learning and machine learning frameworks such as pytorch, tensorflow, Jax, Theano, scikit-learn
- Experience with cloud infrastructure, e.g., AWS, Microsoft Azure, or Google Cloud, and model deployment frameworks such as TorchServe
- Be up to date on deep learning architectures and practices, such as Transformers, RNNs, CNNs, cross validation, language models, self-supervised learning, metric learning, model distillation, and transfer learning
- Write clean, concise, sustainable code and care about software engineering best practices.
- Ability to work with interdisciplinary and geographically distributed teams.
- Strong organizational and communication skills.

**Preferred (but not required)**
- Background knowledge of biology, including DNA, RNA, and proteins
- Masters or other background in bioinformatics or computational biology, especially sequence analysis, with an understanding of core methods like sequence alignment, multiple sequence alignment, PSSMs, phylogenetic trees, Hidden Markov Models, etc. Understanding typical file formats for storing biological sequence and structure data is a plus.
- Experience with large language models in production, e.g., BERT, GTP-3
- Experience with fine-tuning, Bayesian models (e.g., Gaussian processes),
- Experience with active learning, design of experiments, and/or Bayesian optimization.
- Experience working with protein structures and protein structure viewers (e.g., PDB, Chimera, PyMOL).
- Research experience demonstrated by publications in ML conferences, scientific journals, or other peer reviewed venues.