Machine Learning Systems Engineer – Video Computer Vision

Socket.dev; Mount Laurel Township (NJ); On-site; USD 140,000 - 170,000; Full time

Job summary

Socket.dev is seeking a Machine Learning Systems Engineer to build and deploy on-device vision capabilities on Apple devices. You will train and evaluate multimodal vision models, optimize performance, and ensure scalable, low-latency inference within embedded environments.

Ideal candidates have 3+ years in ML, strong Python/C++ coding skills, and experience with PyTorch, SFT pipelines, and end-to-end ML lifecycles for production systems.

Qualifications

• Bachelor's degree in CS or related field with 3+ years of industry experience.
• Strong ML fundamentals and production-grade code for on-device CV/ML features.
• Experience with PyTorch and end-to-end ML lifecycle; SFT pipelines is a plus.

Responsibilities

• Train, evaluate, and deploy on-device vision models on Apple hardware.
• Develop techniques to optimize model performance under tight device constraints.
• Collaborate with hardware, software, and AI teams to ship Apple-quality features.

Skills

Python

C++

PyTorch

Machine Learning

On-device ML

Edge deployment

Education

Bachelor's degree in Computer Science or related

Tools

Swift

CoreML

RealityKit

Job description

The incredible potential of multimodal foundation models and large language models has unlocked machine learning applications that were previously thought infeasible. The Video Computer Vision (VCV) group is looking for a highly motivated and skilled Machine Learning Systems Engineer to help us ship cutting‑edge computer vision technology on Apple devices. The VCV organization has pioneered groundbreaking features like FaceID/FaceKit, Gaze/Hand Gesture Control, Body Tracking, and 2D/3D Scene Understanding fundamentally changing how millions of users interact with technology. We seamlessly balance research and product requirements to deliver pioneering, Apple‑quality experiences. By innovating across the full stack and partnering closely with hardware, software, and AI teams, we shape future products and bring our architectural vision to life.

Description

As a member of the Video Computer Vision team, you will train, evaluate, and deploy purpose‑built vision models on Apple hardware. You will develop innovative techniques to optimize model performance, efficiency, and scalability, ensuring a seamless user experience under strict on‑device constraints.

Minimum Qualifications

Bachelor's degree in Computer Science, Machine Learning, or a related discipline, and 3+ years of relevant industry experience. Strong ML fundamentals. Proven track record of writing high‑quality production code for shipped on‑device CV/ML features deployed on embedded platforms. Solid understanding of operating system fundamentals and extensive programming experience in Python and C++. Hands‑on experience with PyTorch and familiarity with the end‑to‑end ML lifecycle (data preprocessing, training, evaluation, and edge deployment). Experience with Supervised Fine‑Tuning (SFT) pipelines to adapt vision and multimodal foundation models for specialized, on‑device downstream tasks. Robust foundational understanding of machine learning architectures, specifically Multimodal LLMs and the integration of ML components into complex production systems.

Preferred Qualifications

Programming experience with Swift and familiarity with CoreML, CoreFoundation, and RealityKit frameworks. Fundamental knowledge of real‑time video pipelines, image transformations, and rendering loops. Experience optimizing models for neural network accelerators (e.g., Apple Neural Engine or mobile GPUs).