Edge AI Research Scientist — Gen AI & Multimodal Models

Slope; California (MO); On-site; USD 220,000 - 292,000; Full time

Benefits offered by this job

Health benefits

Equity

Competitive compensation

Job summary

Anduril Industries is seeking an AI Research Scientist to serve as a founding ML expert on our Air Dominance & Strike team. You will design, fine‑tune, and deploy next-generation generative AI, LLMs, and agentic systems powering our autonomous air-dominance platforms.

This highly applied research role blends 60% experimentation with 40% hands‑on coding, optimizing models to run on edge hardware and integrating multi‑agent autonomy in critical missions.

Qualifications

• Strong production-level coding in Python and DL frameworks.
• Hands-on experience training, fine-tuning, and evaluating LLMs and multimodal models.
• Advanced degree in CS/ML/Robotics or related field.
• Eligible to obtain and maintain an active U.S. Top Secret security clearance.

Responsibilities

• Develop, pre-train, and fine-tune in-house LLMs and multimodal models.
• Architect and optimize models to run on tactical edge hardware for latency and memory.
• Design and implement robust agentic architectures and multi-agent coordination frameworks.
• Collaborate with computer vision, perception, and motion planning teams to handle multimodal data.
• Define and execute data collection across assets and simulations; build scalable evaluation frameworks.
• Build early-stage prototypes with customers and scale to production features.

Skills

Python

Deep learning

LLM research

Robotics

Education

CS/ML/Robotics PhD

Tools

Axolotl

Hugging Face

DeepSpeed

Megatron-LM

LangChain

LlamaIndex

Job description

Anduril Industries is seeking an AI Research Scientist to serve as a founding ML expert on our Air Dominance & Strike team. You will design, fine‑tune, and deploy next-generation generative AI, LLMs, and agentic systems powering our autonomous air-dominance platforms.

This highly applied research role blends 60% experimentation with 40% hands‑on coding, optimizing models to run on edge hardware and integrating multi‑agent autonomy in critical missions.