Postdoctoral Appointee - AI for Biomedical Discovery
- Compensation
- $72.9K–$121.5K / yr
- Location
- Lemont, IL USA
- Work setup
- Not specified
- Seniority
- Not specified
- Employment
- Contract
- Experience
- 0+ years
- Category
- Not specified
- Posted
Skills
Must have
- Ph.D. completed within the last 0–5 years in computer science, data science, biomedical informatics, computational biology, bioengineering, applied mathematics, electrical engineering, or a related fi
- Strong programming skills in Python and experience developing research or production-quality machine learning software.
- Experience with machine learning or deep learning frameworks such as PyTorch, TensorFlow, JAX, or similar tools.
- Knowledge of federated learning, distributed machine learning, privacy-preserving AI, foundation models, multimodal learning, continual learning, or related areas.
- Ability to design and conduct computational experiments, analyze model performance, and communicate results clearly.
- Experience working with large-scale or complex datasets, including structured, unstructured, multimodal, biomedical, scientific, or high-dimensional data.
- Ability to work independently while contributing effectively to a multidisciplinary research team.
- Strong written and oral communication skills, including the ability to prepare manuscripts, technical reports, presentations, and documentation.
- Ability to model Argonne's core values of impact, safety, respect, integrity, and teamwork.
Nice to have
- Experience developing or extending federated learning frameworks such as APPFL, Flower, FedML, NVIDIA FLARE, or similar systems.
- Experience with multimodal biomedical data, including combinations of clinical records, medical imaging, pathology, genomics, transcriptomics, proteomics, wearable/sensor data, or scientific text.
- Familiarity with foundation models, large language models, vision-language models, biomedical AI models, or model fine-tuning methods such as LoRA, adapters, instruction tuning, or retrieval-augmented
- Experience with continual learning, active learning, reinforcement learning, closed-loop learning, or human-in-the-loop AI workflows.
- Experience with agentic AI frameworks, tool-using LLMs, workflow orchestration, AI planning systems, or multi-agent systems.
- Familiarity with privacy and security techniques such as differential privacy, secure aggregation, secure multiparty computation, homomorphic encryption, trusted execution environments, or secure encl
- Experience with distributed computing, cloud computing, containers, Kubernetes, Docker, Apptainer/Singularity, or high-performance computing environments.
- Experience with MLOps, reproducible workflows, experiment tracking, CI/CD, software testing, benchmarking, or open-source software development.
- Familiarity with biomedical AI validation, data readiness assessment, model evaluation, regulatory-grade evidence generation, or independent verification and validation workflows.
- Demonstrated ability to publish research, contribute to collaborative software projects, or present technical work to interdisciplinary audiences.
Full description
The Argonne team is seeking two highly motivated postdoctoral researchers to help shape the next generation of secure, scalable, and continuously learning AI systems for biomedical discovery. This position will contribute to the Forge project, which is focused on developing advanced multimodal AI capabilities that can learn across distributed data environments without requiring sensitive data to be centralized.
The successful candidates will work at the intersection of federated learning, foundation models, multimodal biomedical AI, privacy-preserving machine learning, continuous learning, and agentic AI systems. This is an opportunity to conduct applied research that advances trustworthy AI for biomedical and national security-relevant use cases while working in a multidisciplinary environment that brings together computer scientists, AI researchers, domain scientists, software engineers, and high-performance computing experts.
You will help design and implement new methods for multimodal federated learning across heterogeneous data types such as clinical, imaging, omics, text, and experimental data. The work will include developing approaches for continual model improvement, adaptive federated training, model evaluation, workflow automation, and AI-assisted orchestration of distributed learning tasks. The position will also provide opportunities to contribute to open-source software, publish research findings, present at major conferences and workshops, and collaborate with partners across national laboratories, universities, government agencies, and biomedical research organizations.
The work will take place in a collaborative, mission-driven research environment that values technical creativity, rigorous engineering, scientific impact, and teamwork. The group works on practical AI systems that connect research prototypes to real-world deployment environments, including cloud, secure enclaves, trusted research environments, and leadership computing platforms. Candidates should be comfortable working in a fast-moving research setting where methods development, software implementation, experimentation, and scientific communication are all important parts of the role.
Core Responsibilities:
- Conduct research and development in federated learning, privacy-preserving machine learning, multimodal AI, and foundation model adaptation for biomedical and related scientific applications.
- Develop new methods for multimodal federated learning that can integrate information across distributed datasets, including imaging, omics, clinical, text, sensor, and other structured or unstructured data modalities.
- Design and implement continuous learning approaches that allow models to improve over time as new data, validation results, or experimental feedback become available.
- Explore agentic AI approaches for federated learning, including AI agents that can assist with task orchestration, experiment planning, model evaluation, workflow automation, and decision support across distributed environments.