Generative AI Engineering Intern

Job Description

Generative AI Engineering Intern | Dassault Systemes

The Tone:
This is a full-time internship at Dassault Systemes, located in San Diego, CA or Durham, NC. The BIOVIA Generative AI Portfolio team works to evolve scientific workflow software by integrating AI agents, Generative AI, and machine learning capabilities. This role is crucial for prototyping and evaluating a generative visual scripting system, enabling scientific users to create and modify workflow pipelines efficiently through structured tool use. It is a hands-on applied AI engineering internship focused on building working systems and learning from real technical constraints.

The TL;DR
• Role: Internship
• Type: Full-time
• Location: San Diego, CA or Durham, NC
• Pay: $22–$32 hourly
• Team: BIOVIA Generative AI Portfolio team
• Mission: Prototype and evaluate a generative visual scripting system for scientific workflows by building and testing AI agents that can create and modify pipelines.
• Tech Stack: Python, LLMs, JSON schemas, REST APIs, Pipeline Pilot, KNIME, Alteryx, Dataiku, MCP framework, RAG, ReAct

What You’ll Actually Do
• Technical Experimentation: Prototype an agent that creates and modifies visual workflows by invoking MCP tools, implementing orchestration logic like tool selection and parameter completion, and building utilities for schema inspection and automated correctness checks.
• Tool Integration: Build and iterate on tools and tool schemas, including defining inputs/outputs, constraints, error handling, and adding guardrails such as allowlists and parameter bounds to ensure predictable agent behavior.
• Evaluation Framework: Define a benchmark set of visual scripting tasks, including creating from intent, editing, refactoring, and debugging, and build an evaluation harness to measure correctness, robustness, and repeatability.
• Expected Outcomes: Deliver a working prototype demonstrating tool-based generative visual scripting in a controlled environment, a repeatable evaluation suite with benchmark scenarios and metrics, and a short report summarizing results and recommendations.

The Must-Haves
• Background: Master’s degree or equivalent experience in Computer Science, Engineering, Data Science, AI/ML. Early Career.
• Experience: Familiarity with LLMs and agent/tool patterns, including function calling, tool routing, and structured outputs; comfort working with JSON schemas, validation, and integration-style development.
• Skills: Strong Python skills, encompassing APIs, data structures, debugging, and packaging; clear written communication for documenting experiments and results.
• Bonus: Experience with workflow/pipeline tools such as Pipeline Pilot, KNIME, Alteryx, or Dataiku; exposure to REST APIs, async execution, or software testing (unit/integration); familiarity with RAG, prompt engineering, ReAct, LLM evaluation methods, or the MCP framework.