Machine Learning Scientist, Open-Endedness

Machine Learning Scientist, Open-Endedness (Level Flexible)San Francisco, CA USAAbout Lila SciencesLila Sciences is the world’s first scientific superintelligence platform and autonomous lab for life, chemistry, and materials science. We are pioneering a new age of boundless discovery by building the capabilities to apply AI to every aspect of the scientific method. We are introducing scientific superintelligence to solve humankind's greatest challenges, enabling scientists to bring forth solutions in human health, climate, and sustainability at a pace and scale never experienced before. Learn more about this mission at www.lila.aiAt Lila, we are uniquely cross-functional and collaborative. We are actively reimagining the way teams work together and communicate. Therefore, we seek individuals with an inclusive mindset and a diversity of thought. Our teams thrive in unstructured and creative environments. All voices are heard because we know that experience comes in many forms, skills are transferable, and passion goes a long way.If this sounds like an environment you’d love to work in, even if you only have some of the experience listed below, please apply.Lila Sciences is seeking experienced, creative, and talented Machine Learning Scientist (Open-Endedness) across Scientist I/II and Senior Scientist levels to join our team. Title will be determined by merit and experience level.Open-Endedness is an emerging area of machine learning that aims to automate never-ending innovative processes of discovery and exploration. The Open-Endedness Team, led by Ken Stanley, investigates in particular how a continual chain of deep transformative creativity can be maintained that far exceeds the derivative creativity seen in current models. In effect, the systems developed on this team will go beyond simply solving problems posed by users, to conceiving the future unimagined directions of science itself.To realize this vision, we’re seeking a broad tapestry of ML expertise to facilitate daring and unconventional investigations, including but not exclusive to pre-training, fine-tuning, RLHF, distillation, mechanistic interpretability, and quality diversity (QD) techniques.What You\'ll Be BuildingDesigning, implementing, and modifying generative models (e.g., LLMs, diffusion models, multimodal models) through unconventional pipelines to achieve unconventional behaviorsUnconventional evaluation techniques, including subjective evaluation and the evaluation of interestingnessCreative approaches to investigating, understanding, and visualizing the internal representations of large models, encompassing mechanistic interpretability but also going in new directions beyond itQuality diversity (QD) algorithms like MAP-Elites, novelty search with local competition, POET, OMNI, minimal criterion novelty search, etc., with LLMs or other large models potentially being updated on the inner loop.What You’ll Need to SucceedPhD in quantitative disciplines ideal, but will consider self-taught researchers with exceptional achievementsPublications in relevant conferences, such as NeurIPS, ICML, AAAI, ICLR, GECCO, ICCCExpertise in ML frameworks (PyTorch/TensorFlow/Jax); and/or QD algorithm implementation; and/or neuroevolution algorithm implementationExperience in training and deploying ML models on distributed computing services (g. AWS/GCP/Azure, or clusters).Lila Sciences is committed to equal employment opportunity regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender identity or Veteran status. #J-18808-Ljbffr