Ground every answer in facts on this page and the original listing. We never invent Glassdoor-style reviews or salaries that are not in our data.
Website: axiombio.com
Public cache only — not an employee review.
Axiombio is hiring an ML Researcher at our San Francisco Global Headquarters to build AI systems that predict drug toxicity in humans—replacing animal testing with machine learning trained on human-relevant experimental data. Role Overview You'll lead machine learning research and engineering for our core product, building models that predict whether drug molecules will harm patients. Starting from raw wet-lab data, you'll design end-to-end systems spanning data cleaning, representation learning, model training, evaluation, deployment, and customer inference. Key Responsibilities Design ML systems that connect molecular chemistry, cellular responses, drug exposure, and human toxicity outcomes. Train multimodal models on chemical structures, cellular images, transcriptomics, proteomics, mass spectrometry, pharmacokinetic data, and clinical outcomes. Build foundational learning systems for molecular structures, biological images, and experimental data across modalities. Conduct error analysis to diagnose model failures and identify high-value data gaps. Collaborate with biologists, chemists, and wet-lab teams to design experiments that improve model performance. Build mechanistic reasoning agents that explain toxicity mechanisms and guide scientific decisions. Drive models from research prototyping through training, evaluation, production deployment, customer feedback, and continuous improvement. Requirements Proven machine learning excellence through research publications, production systems, or substantial personal projects. Skilled in PyTorch or JAX code, experienced in debugging training workflows, managing imperfect data, and constructing production systems. Passion for complex biological modeling problems where data is multimodal, sparse, and scientifically important. Commitment to rigorous evaluation, calibration, and real-world utility over toy benchmarks. Ability to move between research exploration and production engineering. Openness to developing expertise in pharmacology, toxicology, cellular biology, and drug development. Valued experience: representation learning, self-supervised learning, computer vision for microscopy, multimodal learning, distributed training, LLMs, and reasoning systems. About the Company Axiombio is building AI systems to replace animal toxicity testing in drug development. We partner with leading pharmaceutical companies and biotech firms to generate human-relevant experimental datasets, then apply machine learning to predict safety outcomes with greater accuracy than animal models. We currently focus on predicting drug-induced liver injury and count seven of the world's top twenty pharma companies as active users. Our vision: assemble the world's largest human-centered experimental dataset across all organ…
Generated for personal interview prep · 2026-07-17 UTC · getajob.ai