AI Research Internship

Role Overview

Lexsi Labs is a leading frontier lab focused on building aligned, interpretable, and safe Superintelligence. As an AI Research Intern, you will work on large-scale industry problems, pushing the frontiers of AI technologies in a fast-paced startup environment that values curiosity, initiative, and exceptional performance.

Responsibilities

• Architect and enhance open-source Python tooling for alignment, explainability, and robustness.
• Improve model interpretability using SOTA XAI techniques across text, image, and tabular modalities.
• Probe internal model representations and circuits using activation patching and feature visualization.
• Develop and benchmark uncertainty estimation methods and robustness metrics for foundation models.
• Work on Tabular Foundational Model architectures and the TabTune library.
• Explore new algorithms around Reinforcement Learning and fine-tuning libraries.
• Author experiment code, run systematic studies, and co-author whitepapers or conference submissions.

Requirements

• Strong Python expertise in writing clean, modular, and testable code.
• Deep understanding of machine learning and deep learning principles with hands-on experience with PyTorch.
• Comprehensive knowledge of Transformer architectures, including attention mechanisms and models like GPT, LLaMA, and Mamba.
• Proficiency in Git workflows, packaging, and collaborative development.
• Excellent communication skills and a collaborative mindset for peer code reviews.

Skills

• Python
• PyTorch
• Transformers
• Git
• Reinforcement Learning

Nice to Have

• Research publications in venues such as NeurIPS, ICLR, or ICML.
• Prior open-source contributions to AI/ML libraries.
• Experience with model compression, quantization, or pruning.
• Expertise in LLM alignment techniques like RLHF or LoRA.