Robotics AI Engineer

Socket.dev; Cupertino (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Socket.dev in Cupertino is seeking a Robotics AI Engineer to translate user intents into robust, adaptive behaviors and deploy frontier multimodal models on Apple silicon for real-time, interactive experiences.

The role spans prompting, evaluation, distillation, and on-device deployment, requiring hands-on work with transformer and diffusion models, Python, and ML frameworks. A product-driven team emphasizes shipping real, user-facing AI.

Qualifications

• BS or MS in Computer Science or Machine Learning.
• Strong background in modern machine learning methods.
• Experience training or fine-tuning transformer and/or diffusion models.
• Proficiency using coding agents such as Claude Code and Codex.
• Genuine excitement for experimenting with the latest AI models, tools, and multimodal capabilities.

Responsibilities

• Translate natural user intents into robust, adaptive on-device behaviors.
• Work across the full AI/ML stack: prompting, evaluation, distillation, and deployment.
• Ship real products with interactive AI experiences and frontier model integration.

Skills

Transformer models

Diffusion models

Python

PyTorch

TensorFlow

Agent harnesses

Education

BS/MS in CS/ML

PhD in CS/ML

Tools

Claude Code

Codex

llama.cpp

vLLM

MLX

Core ML

Job description

Our team leverages state-of-the-art AI to better understand and communicate with users via semantic and contextual intent understanding. We are looking for a Robotics AI Engineer to bring frontier reasoning and multimodal models to life in real-time, interactive experiences.

Description

In this role, you will translate natural user intents into robust, adaptive behaviors. You will work across the full stack of applied AI/ML: prompting and evaluating frontier models, distilling them into efficient on-device models, and deploying them on Apple silicon. A successful candidate is energized by rapidly experimenting with the latest AI models, tools, and agent harnesses, and has a strong foundation in modern machine learning with hands-on experience training and fine-tuning their own models. Above all, this is a deeply product-driven team that values engineers who have shipped real products and are motivated by putting great experiences into users' hands.

Minimum Qualifications

• BS or MS in Computer Science or Machine Learning
• Strong background in and deep understanding of modern machine learning methods
• Experience training or fine-tuning your own transformer and/or diffusion models
• Proficiency using coding agents such as Claude Code and Codex
• Genuine excitement for experimenting with the latest AI models, tools, and multimodal capabilities

Preferred Qualifications

• PhD in Computer Science, Machine Learning, or a related field
• Experience with model fine-tuning and distillation for efficient, on-device deployment
• Hands-on experience with inference engines such as llama.cpp, vLLM, MLX, or Core ML
• Experience building agent harnesses for multi-step, closed-loop reasoning
• Track record of shipping products to users
• Proficiency in Python and ML frameworks (PyTorch or TensorFlow)