Position Overview
The construction industry, a $15 trillion global sector, is increasingly shaped by AI, automation, and data-driven insights. We are seeking a motivated and innovative Research Intern to develop and prototype multimodal AI agents that can interpret and reason across diverse Architecture, Engineering, and Construction (AEC) data, designing creative new workflows for AI in Construction.
This internship will focus on Large Language Models (LLMs), Vision-Language Models (VLMs), and knowledge graphs to create intelligent agents capable of understanding AEC documentation, images, and sensor data to autonomously execute construction management tasks.
As an intern with Autodesk Research, you’ll collaborate with a global team working at the intersection of AI, robotics, and construction technology. Your work will directly contribute to advancing digital twins, multimodal reasoning, and scalable AI systems for the AEC industry.
Responsibilities
- Multimodal AI Agent Development: Design and prototype AI agents using LLMs, vision-language models, and multi-agent frameworks. Focus on integrating agent protocols to enable robust interactions between agents and external tools (e.g., APIs, databases, and resources).
- Prototype AI Solutions in Construction Project: Build a functional AI agent prototype, including demos that showcase capabilities in querying, data analysis, and inter-agent interactions in a real-world AEC project.
- Research and Implement Data Analysis Techniques: Explore graph querying methods and multimodal data to interpret and combine textual, visual, and sensor data from construction sources. Address key challenges, including handling domain-specific nuances in AEC data.
- Conduct Experiments and Present Findings: Perform hands-on research on AI agent architectures for construction applications and present results to the team through reports, demos, or potential contributions to academic papers.
Minimum Qualifications
- Pursuing a Master's or PhD in Computer Science, Machine Learning/AI, Robotics, or a related field, with familiarity in AEC domains.
- Strong proficiency in Python and experience with AI frameworks such as LangChain, PyTorch, or similar for building agents and models.
- Hands-on experience with LLMs, vision-language models (e.g., GPT-4V, Claude, Gemini), and graph tools (e.g., Neo4j, GraphQL, or similar).
- Understanding of multimodal data processing, including techniques for combining textual, visual, and sensor inputs.
- Solid grasp of agent protocols (e.g., Model Context Protocol, Agent2Agent Protocol) and experience with external tool integrations (e.g., APIs).
- Excellent analytical, problem-solving, and communication skills, with a passion for innovative technology in construction.
