Postdoctoral Researcher — AI for Scientific Reasoning and Hypothesis Generation

Position Overview Dr. Avik Mahata's group at Merrimack College is seeking a highly motivated postdoctoral researcher in computer science, machine learning, or related fields to develop next-generation AI systems for scientific reasoning and autonomous hypothesis generation . The project focuses on building agentic AI systems that integrate large language models, knowledge graphs, and external computational tools into a unified framework capable of: reasoning over multimodal scientific data, identifying gaps in knowledge, and generating and prioritizing testable hypotheses. While the application domain includes materials science and experimental data, the primary research contribution is in AI systems, machine learning, and computational reasoning architectures . Research Scope The postdoc will lead the design and development of a tool-augmented, multi-agent AI system , including: Agentic LLM architectures for structured reasoning and planning Knowledge graph construction and querying (e.g., Neo4j, graph embeddings) Multimodal representation learning across text, images, and simulation data Tool integration frameworks (function-calling, APIs, distributed systems) Uncertainty-aware reasoning and hypothesis scoring Verification and validation pipelines for scientific inference This work sits at the intersection of: machine learning and AI systems scientific computing and data integration automated reasoning and decision-making Required Qualifications PhD in Computer Science, Machine Learning, Artificial Intelligence, Data Science , or closely related field Strong programming skills (Python required) Experience with modern ML frameworks (e.g., PyTorch, JAX, TensorFlow) Background in one or more of the following: large language models (LLMs) or generative AI knowledge graphs or graph machine learning multimodal learning distributed or systems-level AI Preferred Qualifications Experience with agent-based AI systems or tool-augmented LLMs Familiarity with scientific computing workflows or simulation tools Experience with uncertainty quantification, probabilistic modeling, or reasoning systems Exposure to materials science, physics, or chemistry (helpful but not required) What We Offer Opportunity to lead a cutting-edge research effort in AI for scientific discovery Collaboration across computer science, engineering, and physical sciences Access to HPC resources and multimodal scientific datasets Opportunities for high-impact publications and grant development (NSF/DOE) Mentorship and support for career advancement in academia or industry How to Apply Please send the following materials to mahataa@merrimack.edu : CV Brief research statement (1-2 pages) Representative publications or code (GitHub link encouraged)

Location:

Merrimack College, North Andover, MA PI:

Dr.

Avik Mahata Start Date:

Fall 2026

Duration:

1 (renewable) Additional Information This position is part of a broader effort to develop generalizable AI systems for scientific reasoning , aligned with national initiatives in AI for science. The goal is to move beyond data-driven pattern recognition toward autonomous, verifiable, and interpretable scientific discovery systems .

Pay:

$65,000.00 - $80,000.00 per year

Work Location:

Hybrid remote in North Andover, MA 01845