Principal Machine Learning Scientist

Experience working with text data to build Deep Learning and ML models, both supervised and unsupervised. Strong understanding of the math and theory behind machine learning and deep learning. Software engineering background with at least 3-5 years of experience using Python, SQL, Unix-based systems, git, and github. Machine / Deep Learning development skills, including experiment tracking (AWS SageMaker, Hugging Face, transformers, PyTorch, scikit-learn, Jupyter, Weights & Biases). Understanding of Language Models, using and training / fine-tuning, and familiarity with industry-standard LM families. Master's degree or PhD in Computer Science, Electrical Engineering, AI, Machine Learning, applied math or related field, with relevant industry experience, or outstanding previous achievements. Excellent communication and teamwork skills. Fluent in written and spoken English. Familiarity in coding for at-scale production, building back-end API services or stand-alone libraries. Essential dev-ops skills (Docker, AWS EC2/Batch/Lambda). Familiarity in building front-ends (LLMs or more standard React, Javascript, Flask) for demos, POCs and prototypes. Experience with advanced prompting, fine-tuning or training an LLM, open-source or cloud. Showcase previous work (e.g., via a website, presentation, open source code).

Work with subject matter experts and product owners to determine questions to be asked and answered. Curate, generate, and annotate data, and create optimal datasets. Answer questions and make trainable datasets from raw data using SQL and scripting languages. Develop and tune Machine Learning models. Utilize, adopt, and fine-tune Language Models, including third-party LLMs and locally hosted LMs. Stay current in the field by reading research papers and experimenting with new architectures and LLMs. Optimize models for scaled production usage. Communicate insights, behavior, and limitations of models to peers, subject matter experts, and product owners. Write clean, efficient, and modular code with automated tests and documentation. Stay up to date with technology and make explainable technological choices.