Full-Time

Machine Learning / Perception Engineer

Posted on 9/24/2025

Swarmbotics AI

Swarmbotics AI

11-50 employees

Low-cost autonomous ground swarm robotics platform

No salary listed

Phoenix, AZ, USA

In Person

Relocation to Phoenix, AZ area required; export-controlled work eligibility.

Category
AI & Machine Learning (1)
Required Skills
Python
Pytorch
NumPy
C/C++
OpenCV
Requirements
  • Minimum 2 years industry experience in ML related field
  • Bachelor's degree in computer science, math, or related field
  • Strong mathematical foundation of neural networks, transformers, convolutional networks, and statistics
  • Python, PyTorch, C++, ROS, ONNX, TensorRT
  • Familiarity with modern ML / computer vision tools such as OpenCV, Numpy, and Pillow
  • Strong organization and communication skills to work well across teams in a fast-paced startup environment
  • Comfort working in the high-paced, fluid environment of a tech startup
  • Excitement about contributing to the defense of the United States and its allies
  • Must be eligible to work on export-controlled projects
  • Ability to relocate to Phoenix, AZ area
Responsibilities
  • Design, implement, train, and test multiple deep learning models including: Deep computer vision models
  • Dense disparity prediction models
  • Reinforcement learning models
  • Vision Language Action (VLA) models
  • Object detection and panoptic segmentation models
  • Optimize models to run real-time on edge devices (Jetson Orin) through model quantization and distillation
  • Deploy trained models in the field
  • Write robust Python, PyTorch, and C++ code
  • Work collaboratively across multiple teams such as SLAM/localization, navigation, tracking, and swarm intelligence
Desired Qualifications
  • Masters or PhD in computer science, math, engineering, or related field
  • Prior robotics experience or autonomous driving background
  • Experience with owning an entire data to production model pipeline
  • Academic publications in related fields

Swarmbotics AI develops low-cost, ground swarm robotics for defense and industrial use. Its ANTS (Attritable, Networked, Tactical Swarm) ecosystem centers on unmanned ground vehicles that operate autonomously in swarms. The core product is a modular platform that can be adapted for many missions, enabling scalable, unsupervised autonomous operation in dangerous or large-scale environments. Compared to competitors, Swarmbotics AI emphasizes affordability and modularity to support wide deployment, offering attritable technology to reduce risk and cost while enabling large numbers of robots to work together. Its goal is to make swarm robotics a practical reality by providing scalable, cost-effective autonomous UGV swarms that can be used across defense and industrial applications.

Company Size

11-50

Company Stage

Seed

Total Funding

$17M

Headquarters

Phoenix, Arizona

Founded

2023

Simplify Jobs

Simplify's Take

What believers are saying

  • U.S. Army awarded contract February 2025 for 1st Cavalry Division integration.
  • xTech Overwatch success accelerates 2026 deployment of hundreds of robots.
  • Founders' AWS and Marine Corps expertise drives cost-effective autonomy.

What critics are saying

  • Anduril captures Army swarm contracts with superior funding in 12 months.
  • Shield AI's Hivemind obsoletes Swarmbotics autonomy in 6 months.
  • Ukraine UGV failures in mud erode 1st Cavalry trust by 2027.

What makes Swarmbotics AI unique

  • ANTS ecosystem features modular Haul ANT and Fire ANT for heterogeneous swarms.
  • AntsNet software enables hardware-agnostic, real-time swarm coordination.
  • FireAnt UGVs deliver anti-tank swarm tactics at low cost.

Help us improve and share your feedback! Did you find this helpful?

Benefits

Flexible Work Hours

Growth & Insights and Company News

Headcount

6 month growth

-5%

1 year growth

-2%

2 year growth

-2%
Defence Blog
Feb 5th, 2026
U.S. Army selects Swarmbotics AI for autonomous robot swarms

U.S. Army selects Swarmbotics AI for autonomous robot swarms. * The U.S. Army selected Swarmbotics AI to develop swarming, attritable small unmanned ground vehicles for Transformation in Contact with the 1st Cavalry Division following xTech Overwatch evaluations. * The program aims to field large numbers of autonomous ground robots to support maneuver units and accelerate operational integration of ground autonomy. The United States Army has selected Swarmbotics AI to develop swarming, attritable small unmanned ground vehicles as part of its Transformation in Contact effort with the 1st Cavalry Division, the company announced on Feb. 5. The award follows Swarmbotics AI's performance at last year's xTech Overwatch competition, where Army soldiers evaluated autonomous systems intended for rapid integration into operational units. The effort is focused on advancing the Army's ability to deploy large numbers of low-cost robotic platforms to support maneuver formations in contested environments. According to the company, the program centers on swarming, heterogeneous small unmanned ground vehicles designed to operate cooperatively and accept attrition during combat operations. Swarmbotics AI said the systems are intended to create tactical dilemmas for adversaries while reducing reliance on more expensive, crewed platforms. "Mass is our objective, by employing swarms of heterogeneous small sUGVs we create multiple dilemmas for our adversaries at fractions of the cost of exquisite platforms," said Stephen Houghton, chief executive officer of Swarmbotics AI, in a company statement. During xTech Overwatch evaluations, soldiers from the 1st Cavalry Division assessed multiple contractors' autonomous capabilities across ground vehicles, airborne systems, payload integration, and mission planning tools. The assessments were conducted in coordination with Army stakeholders to determine which technologies were suitable for rapid transition into Transformation in Contact formations. Lt. Col. Nick Rinaldi, competition lead at the Army Applications Laboratory, said the structure of xTech Overwatch accelerated decision-making by placing operational units directly in the evaluation process. "Having the entire material kill chain as evaluators invested in the outcome was critical to the speed of decisions on which technologies were most suited to move into a TiC formation," Rinaldi said. He added that the approach reflects how the Army intends to work with industry going forward, emphasizing rapid experimentation, direct soldier feedback, and quicker transition from demonstration to field use. According to Rinaldi, engaging companies on real operational problems and evaluating systems in realistic contexts allows the Army to make procurement and integration decisions faster. Swarmbotics AI's technical leadership said the program demonstrated the operational value of robotic mass at the small-unit level. Drew Watson, the company's chief technology officer, said xTech Overwatch showed what could be achieved with limited numbers of robots and pointed to larger-scale deployments planned with the 1st Cavalry Division. "The key takeaway is that robot swarms are here to stay," Watson said. He added that upcoming work with the division in 2026 will focus on establishing baseline concepts for deploying hundreds of robots during maneuver and breaching operations. The Army's Transformation in Contact initiative is designed to push new capabilities directly into operational units for experimentation and refinement, rather than relying solely on long-term development programs. Attritable unmanned ground systems are viewed as a way to extend reconnaissance, security, and support functions while reducing risk to soldiers. Readers who wish to follow its weekly coverage can subscribe to the Weekly Defense Roundup. If you wish to report a grammatical or factual error in this article, please let Defence Blog know by using the online form.

TechCrunch
Aug 19th, 2024
Swarmbotics founders grew ‘obsessed with robot swarms’ and now plan to bring them to the battlefield

Swarmbotics AI is developing autonomous ground robots that can make coordinated decisions to generate a real edge for the warfighter.

INACTIVE