Machine Learning Scientist, AI Explainability

Description

• Lead the Prometheus team’s mission to make AI-driven battery science transparent and trustworthy. You will design, train, and deploy interpretable Large Language Models and multi-agent systems that explain how AI-discovered electrolyte chemistries translate into record-breaking Li-Metal and Li-ion performance.
• Own the full research lifecycle—from literature mining and dataset curation to model architecture, training, evaluation, and deployment—ensuring every insight is reproducible, auditable, and aligned with SES’s safety and sustainability standards.
• Pioneer novel explainability techniques (attention visualizations, concept activation vectors, counterfactual generators, causal graphs) that allow electrochemists, cell engineers, and regulatory bodies to understand why an AI-recommended additive stabilizes the SEI layer or suppresses dendrite growth.
• Build interactive dashboards and APIs that surface model rationales in real time; integrate them into SES’s internal experimentation platform so scientists can query, “Why did the model suggest this solvent ratio?” and receive human-readable evidence plus statistical confidence scores.
• Collaborate cross-functionally with battery modeling, robotics, and cloud infrastructure teams to embed interpretability hooks at every layer—from high-throughput robotic synthesis logs to cloud-scale molecular dynamics simulations—creating a seamless feedback loop between lab data and model explanations.
• Establish rigorous benchmarking protocols and open-source evaluation suites that set the industry standard for AI explainability in materials science; publish findings at NeurIPS, ICML, and domain-specific journals such as Nature Energy or Joule.
• Mentor junior researchers and PhD interns, fostering a culture of scientific rigor, ethical AI, and fearless experimentation; host weekly reading groups on interpretability, fairness, and robustness to accelerate collective learning.
• Translate complex technical insights into executive-level narratives that guide strategic decisions on which chemistries to scale, which patents to file, and which partnerships to pursue—directly influencing SES’s multi-gigawatt-hour pipeline.
• Champion responsible AI governance by working with legal, compliance, and ESG teams to ensure our models meet emerging global regulations (EU AI Act, SEC climate disclosures, DOE critical-material guidelines) while maintaining competitive advantage.
• Contribute to SES’s broader AI-for-science roadmap, identifying adjacent opportunities where explainable models can unlock new physics-informed discoveries in solid-state batteries, sodium-ion chemistries, and next-generation cathode coatings.

Requirements

• PhD or MS in Computer Science, Machine Learning, Statistics, or related quantitative field with 3+ years of post-graduate experience developing interpretable or explainable AI systems.
• Demonstrated expertise in transformer architectures, attention mechanisms, and large-scale pre-training; hands-on experience with PyTorch, JAX, or TensorFlow in multi-GPU or TPU environments.
• Proven publication record (NeurIPS, ICML, ICLR, or top-tier domain journals) in explainability, causal inference, or trustworthy ML applied to scientific data.
• Strong software engineering practices: version control (Git), containerization (Docker), CI/CD, and cloud platforms (AWS, GCP, or Azure); ability to productionize models at scale.
• Nice-to-have: domain knowledge in chemistry, materials science, or battery technology; experience with graph neural networks, molecular representations (SMILES, SELFIES, 3D point clouds), or lab automation datasets.

️ Benefits

• Fully remote-first culture with flexible hours and asynchronous collaboration, plus quarterly on-site summits in Boston or Singapore to align with lab teams.
• Competitive equity package (NYSE: SES) and performance bonus tied to breakthrough milestones in battery energy density and cycle life.
• Annual $5,000 professional development stipend for conferences, courses, or certifications; dedicated 20 % “innovation time” to pursue blue-sky research.
• Comprehensive health, dental, vision, and mental-wellness coverage for employees and dependents, plus 12 weeks of gender-neutral parental leave.

Apply tot his job Apply To this Job