Apply on Employer Site

X, The Moonshot Factory · 2 weeks ago

2026 PhD Residency - Knowledge Graph, Operations Research and Reinforcement Learning, Early Stage Project

Mountain View, CA

Full-time

Onsite

New Grad

$109K/yr - $150K/yr

X, The Moonshot Factory is seeking a PhD resident specializing in machine learning and artificial intelligence. The role involves collaborating with project teams to develop algorithms and solutions in reinforcement learning while engaging with the resident community and participating in tech talks.

AerospaceAugmented RealityAutomotiveRobotics

Responsibilities

Collaborating with the X Project team to design and develop machine learning approaches/algorithms to support the project's mission

Understanding the project’s goals and challenges and conduct relevant literature surveys

Suggesting agentic RL solutions in which machines/approaches/algorithms/models can apply to those challenges

Prototype/train/evaluate RL algorithms, tools and components of large scale AI systems

Engaging with the resident community during the program, engaging with the X community, attending colloquia and tech talks

Qualification

Reinforcement LearningLarge Language ModelsPythonOperations ResearchKnowledge GraphsC/C++Synthetic Data GenerationOpen-source ProjectsGraph Query Languages

Required

Currently enrolled in a PhD program in a STEM field such as CS, Physics, Engineering or Mathematics/Statistics with a strong interest in machine learning

Strong experience with one or more general purpose programming languages such as Python, C/C++

Experience with reinforcement learning (RL) and large language models (LLMs)

Ability to set up RL infrastructure and synthetic data generation pipelines

Familiarity with operations research (OR) problems and their formalization

Knowledge of basic ontologies and programmatically constructing knowledge graphs (KG) from structured and unstructured data

Preferred

Open-source projects that demonstrate relevant skills and/or publications in relevant conferences and journals (e.g. NeurIPS, ICML, ICLR, CVPR, ICCV, COLM, ACL/EMNLP, ICASSP)

Experience with GRPO/DAPO based online reinforcement learning with LLMs(7B+) on multi-GPU settings

Ability to produce code (e.g., in Google OR-Tools) and define rewards for formalized optimization problems

Experience translating natural language into formal graph query languages (Cypher, SPARQL)

Experience using frontier LLMs (e.g., Gemini-Pro) to generate large-scale, high-quality synthetic datasets with automated verification steps