Research Engineer

Acceler8 Talent; Mountain View (CA); On-site; USD 140,000 - 210,000; Full time

Job summary

Acceler8 Talent in Mountain View, CA seeks a Research Engineer to help build an AI-powered observability platform for critical infrastructure. You will work on multimodal AI systems that combine language, vision, and agentic reasoning to monitor large physical environments.

The role emphasizes production-grade software and real-world deployment in a fast-moving startup environment, collaborating with robotics and autonomy engineers to scale the platform for diverse industrial settings.

Qualifications

• Experience with vision-language models or multimodal systems.
• Strong background in ML or applied AI research.
• Experience taking research concepts into real-world AI deployments.
• Ability to thrive in autonomous startup environments.
• Production-grade software development skills.
• Familiarity with robotics or embodied AI is a plus.

Responsibilities

• Build multimodal AI systems combining language, vision, and agentic reasoning.
• Develop VLMs and multimodal agents to analyze and act on complex data.
• Design systems for natural interaction with fleets of autonomous robots.
• Work on large-scale reasoning systems to analyze incidents and identify faults.
• Build infrastructure grounding language models in visual observations from drones/robots.
• Develop retrieval, memory, and decision-making for mission-critical environments.
• Collaborate with robotics engineers to bridge perception, reasoning, and execution.

Skills

Vision-Language Models

Multimodal AI

LLMs

Agentic systems

Retrieval systems

Robotics perception

Software engineering

Real-world deployment

Job description

I am seeking a Research Engineer to join a venture-backed AI and robotics startup building an AI-powered observability platform for critical infrastructure. This is an opportunity to work on cutting-edge multimodal AI systems that combine language, vision, agentic reasoning, and robotics to monitor and manage some of the world's largest physical environments, including data centers, solar farms, wind farms, and oil refineries.

The company has already deployed its technology with customers, recently secured FAA approval for autonomous drone operations, and is founded by former Google and Meta robotics engineers with multiple successful exits valued at over $450M combined.

What You’ll Do:

• Build multimodal AI systems combining language models, vision models, and agentic reasoning
• Develop VLMs and multimodal agents capable of understanding, analyzing, and acting on complex real-world data
• Design systems that allow users to interact naturally with fleets of autonomous robots
• Work on large-scale reasoning systems that analyze historical incidents, identify faults, and recommend corrective actions
• Build infrastructure for grounding language models in visual observations from drones and robotic platforms
• Develop retrieval, memory, and decision-making systems for mission-critical environments
• Collaborate closely with robotics and autonomy engineers to bridge perception, reasoning, and execution
• Help shape the future architecture of a rapidly evolving AI platform operating in the physical world

What We’re Looking For:

• Experience working with Vision Language Models (VLMs), multimodal foundation models, or Vision Language Action (VLA) systems
• Strong background in machine learning, deep learning, or applied AI research
• Experience with LLMs, agentic systems, retrieval, or multimodal reasoning
• Exposure to computer vision, perception, robotics, autonomy, or embodied AI is highly desirable
• Strong software engineering skills with the ability to build production-grade AI systems
• Experience taking research concepts into real-world applications
• Ability to thrive in highly autonomous, fast-moving startup environments

This role sits at the intersection of some of the most exciting areas in AI today: multimodal models, agentic systems, robotics, and real-world deployment.

You’ll help build an intelligent platform capable of identifying faults across infrastructure the size of small cities, analyzing historical and live operational data, coordinating robotic inspections, and guiding human operators towards the fastest and most effective resolution.