ML/AI Engineer | Anduril Global Negotiation Guide

Negotiation DNA: Growth-Stage Defense Tech Equity + Mission Premium | $14B+ Private Valuation | Autonomous Systems/Computer Vision/Decision AI | High-Intensity Culture

Region	Base Salary	Equity (Private/4yr)	Bonus	Total Comp
Costa Mesa, CA	$170K-$230K	$120K-$330K	10-15%	$235K-$410K
Seattle, WA	$170K-$230K	$120K-$330K	10-15%	$235K-$410K

Negotiation DNA

ML/AI Engineers at Anduril build the autonomous intelligence that powers defense systems: computer vision for threat detection, reinforcement learning for autonomous navigation, multi-agent coordination for drone swarms, and decision AI for the Lattice platform. Your models operate in adversarial environments where the "data distribution" is an enemy actively trying to defeat your algorithms -- jamming sensors, deploying decoys, and exploiting ML weaknesses. ML deployment at Anduril has lethal consequences: your target classification model determines what gets engaged. This is the highest-stakes ML engineering in the world. Anduril competes directly with OpenAI, Google DeepMind, and top AI labs for ML talent, making this one of the hardest-to-fill roles in the company. [Source: Anduril ML Comp 2025-2026]

Level Mapping: Anduril ML/AI Eng = Google L4-L5 ML = DeepMind Research Eng -- with adversarial deployment, edge inference, and lethal consequence

Culture & Intensity Negotiation

• Adversarial ML Deployment: "My models operate against active adversaries -- enemies jamming sensors and deploying decoys. This adversarial deployment is orders of magnitude harder than commercial ML."
• Lethal Consequence ML: "My target classification models determine engagement decisions. ML errors have lethal consequences -- not user experience degradation."
• ML Talent Competition: "Anduril competes with OpenAI, DeepMind, and top AI labs for ML talent. My willingness to apply ML to defense at this compensation level is valuable."

Global Levers

1. Adversarial ML -- Hardest Deployment Environment: "My models must be robust against active adversaries exploiting ML weaknesses. This is the hardest ML deployment environment in the world."
2. Autonomous Systems ML: "I build perception, planning, and control for autonomous defense systems. Multi-agent coordination for drone swarms is among the most complex ML problems."
3. Lethal AI Premium: "My ML models have lethal consequences -- the highest-stakes AI deployment in the industry. This premium must be reflected in compensation."
4. ML Talent Scarcity in Defense: "ML engineers willing to apply skills to defense are rare. Competing offers from [OpenAI/Google/Anthropic] are significantly higher."

Negotiate Up Strategy: "I build ML systems deployed against active adversaries with lethal consequences -- the hardest and highest-stakes ML in the world. ML engineers willing to work in defense are rare, and my competing offers from [OpenAI/DeepMind] are $400K+ total comp with liquidity. I'm targeting $310K in equity over 4 years. Anduril's autonomous systems mission is unique -- but the equity must reflect both the ML scarcity premium and the adversarial deployment difficulty." Accept at $265K+ equity.

Evidence & Sources

• [Anduril ML/AI Comp -- Levels.fyi 2025-2026]
• [Adversarial ML Deployment -- Defense AI Robustness]
• [Autonomous Systems -- Multi-Agent Drone Swarm Coordination]
• [ML Talent Competition -- Defense vs. Commercial AI Labs]