This role pays more than
74%
of similar roles. Most pay
$187,000–$256,762
— the shaded band above.
At the midpoint, this role pays about
$256k
versus about
$222k
for comparable roles.
Based on 240 similar postings.
Employer
About Anduril Industries
Anduril Industries is a defense technology company that builds advanced hardware and software systems for national security, including autonomous drones, surveillance systems, and the Lattice AI command platform.
Anduril Industries currently has
1697 open roles
on FindRole.
Listed pay typically runs
$146,000–$194,000
across 1504 roles with salary data.
Staff Gen AI Research Scientist
The Staff Gen AI Research Scientist joins the Air Dominance & Strike team to serve as a founding expert on generative AI, LLMs, and agentic systems for air-dominance platforms. This highly applied role involves designing, fine-tuning, and deploying models that power collaborative autonomous behaviors across multi-domain missions. You will develop in-house LLMs and multimodal foundation models using SFT, RLHF, and DPO techniques while optimizing them for power-constrained edge hardware through quantization and pruning. Key responsibilities include building agentic architectures, defining data collection strategies, and collaborating with perception and motion planning teams to integrate camera feeds, radar, and telemetry. Required skills include Python, PyTorch or JAX, and experience with tools like Axolotl, Hugging Face, DeepSpeed, LangChain, or LlamaIndex. The role addresses the challenge of creating autonomous systems for complex military operations in high-stakes environments.
Develop, pre-train, and fine-tune in-house LLMs and multimodal foundation models using SFT, RLHF, and DPO techniques.
Optimize model architecture for deployment on power-constrained edge hardware through quantization, distillation, and pruning.
Design and implement agentic architectures and multi-agent coordination frameworks for complex military missions.
Integrate diverse data modalities including camera feeds, radar, telemetry, and text-based orders into reasoning systems.
Define and execute data collection strategies across physical assets, test sites, and virtual simulations.
Build scalable evaluation frameworks to measure model performance, reliability, and safety in high-stakes environments.
Develop early-stage prototypes based on customer feedback and scale them into production-grade features.
What we're looking for
Production-level coding skills in Python and deep learning frameworks like PyTorch or JAX.
Hands-on experience training, fine-tuning, and evaluating LLMs, Generative AI, or multimodal models.
A background in a technical discipline with 2+ years of experience focusing on generative models and modern transformer architectures.
Experience using modern model training, alignment, and orchestration tools such as Axolotl, Hugging Face, DeepSpeed, Megatron-LM, LangChain, or LlamaIndex.
Ability to move from ambiguous requirements to functional prototypes in a fast-paced environment.
A degree (B.S., M.S., or Ph.D.) in Computer Science, Machine Learning, Robotics, Physics, Mathematics, or a related technical field.
Eligible to obtain and maintain an active U.S. Top Secret security clearance.
Experience with edge accelerators, RLHF/DPO loops, multimodal architectures, or shipping AI products (preferred).