Principal Applied Scientist- AI

Life at Ui Path

The people at UiPath believe in the transformative power of automation to change how the world works. We’re committed to creating category-leading enterprise software that unleashes that power.

To make that happen, we need people who are curious, self-propelled, generous, and genuine. People who love being part of a fast-moving, fast-thinking growth company. And people who care—about each other, about UiPath, and about our larger purpose.

Could that be you?

At UiPath, we believe the next generation of software won’t simply automate tasks—it will reason, plan, adapt, and act autonomously. As a Principal Applied Scientist, you’ll help invent that future. You will lead the research and development of frontier AI systems that combine foundation models, multimodal reasoning, agentic workflows, and post-training techniques to build autonomous software capable of operating enterprise applications like a human.

This is an individual contributor leadership role for scientists who enjoy building. You’ll work at the intersection of applied research and production engineering—designing new algorithms, training and evaluating large models, developing novel post-training methods, and shipping systems that reach enterprise customers. If you’re passionate about pushing the limits of LLMs, multimodal models, reasoning, planning, tool use, and AI agents—and you enjoy transforming research into products—this is the role for you.

Our Advanced Machine Learning organization is building large multimodal foundation models capable of understanding enterprise software, reasoning over complex workflows, planning long-horizon tasks, and autonomously interacting with digital systems.

• Foundation models
• Agentic AI
• Multimodal reasoning
• Reinforcement learning
• Post-training
• Model alignment
• Evaluation systems
• Enterprise AI agents

Our mission is to advance practical general intelligence for enterprise automation.

• Lead research and development of large-scale foundation models and agentic AI systems for enterprise automation.
• Design novel approaches for post-training, model alignment, reinforcement learning, preference optimization, and synthetic data generation.
• Develop scalable evaluation frameworks that measure reasoning, tool use, planning, and autonomous task completion.
• Build production-quality AI systems that combine LLMs, multimodal models, retrieval, planning, memory, and external tool use.
• Drive innovations in long-context reasoning, workflow generation, autonomous software interaction, and multi-agent orchestration.
• Design large-scale experiments, analyze model behavior, and iterate using data-driven evaluation.
• Collaborate closely with engineering and product teams to transition research into production systems.
• Mentor scientists and engineers while raising the technical bar across the organization.
• Influence UiPath’s long-term AI strategy through technical leadership and scientific innovation.

We’re looking for scientists who have successfully taken frontier AI research from idea to production.

• Foundation model training
• LLM post-training
• Reinforcement Learning from Human Feedback (RLHF)
• Preference optimization (DPO, PPO, GRPO or similar methods)
• Reward modeling
• Synthetic data generation
• Model evaluation
• Agentic reasoning
• Tool use
• Function calling
• Long-horizon planning
• Multimodal learning
• Retrieval-Augmented Generation (RAG)
• Distributed training and inference
• Large-scale PyTorch systems
• Model alignment and safety

• PhD or equivalent research experience in Computer Science, Machine Learning, Artificial Intelligence, Statistics, Robotics, or a related discipline.
• Significant experience building and deploying production AI systems based on large language models or multimodal foundation models.
• Demonstrated ownership of end-to-end ML systems—from research and experimentation through production deployment.
• Strong programming skills in Python and deep experience with modern ML frameworks such as PyTorch.
• Experience training or fine-tuning large models using distributed compute infrastructure.
• Strong publication record or demonstrated technical leadership…