AI Video Engineer Intern

E1; United States; Remote; USD 28,000 - 39,000; Full time

Job summary

E1 is seeking an ML intern to help build end-to-end video processing pipelines and tooling for ingesting, preprocessing, and evaluating creator footage to support downstream verification systems. You will work on real-world video data and contribute to prototype-to-production iterations.

You'll apply computer vision and multimodal models, benchmark performance, and document failure modes with rigor, while learning in a hands-on environment that emphasizes rapid iteration and production readiness.

Qualifications

• Currently studying computer science or machine learning, or building equivalent skills through self-directed learning, bootcamps, or substantial side projects.
• Working proficiency in Python and familiarity with the broader ML ecosystem (common frameworks, data tooling, training and evaluation workflows).
• Coursework, research, or project experience in computer vision, video processing, or multimodal machine learning.
• Disciplined experimental habits, including the ability to design fair comparisons and distinguish genuine improvements from random variance.

Responsibilities

• Design and build pipelines that ingest raw video content and prepare it for downstream model evaluation.
• Apply and benchmark vision and multimodal models against real creator footage to assess performance in production-like conditions.
• Track and report model quality with rigor, including clear documentation of failure modes and edge cases.
• Move quickly through prototyping phases, then harden the strongest approaches into reliable, production-ready code.

Skills

Python

ML tooling

Computer vision

Multimodal ML

Experimentation

Education

Pursuing CS/ML

Tools

PyTorch

TensorFlow

OpenCV

Job description

Role overview

Join an engineering team where video is the central medium and verifying its authenticity is mission-critical. As an intern, you will help build the tooling that ingests, processes, and analyzes creator footage so that downstream verification pipelines can confidently assess what they receive. This is a hands-on applied ML role combining computer vision, multimodal modeling, and rapid iteration from prototype to production.

Responsibilities

• Design and build pipelines that ingest raw video content and prepare it for downstream model evaluation.
• Apply and benchmark vision and multimodal models against real creator footage to assess performance in production-like conditions.
• Track and report model quality with rigor, including clear documentation of failure modes and edge cases.
• Move quickly through prototyping phases, then harden the strongest approaches into reliable, production-ready code.

Requirements

• Currently studying computer science or machine learning, or building equivalent skills through self-directed learning, bootcamps, or substantial side projects.
• Working proficiency in Python and familiarity with the broader ML ecosystem (common frameworks, data tooling, training and evaluation workflows).
• Coursework, research, or project experience in computer vision, video processing, or multimodal machine learning.
• Disciplined experimental habits, including the ability to design fair comparisons and distinguish genuine improvements from random variance.

Nice to have

• Exposure to large multimodal models that combine video, audio, or text understanding.
• Experience shipping research code toward production systems, including performance and cost considerations.
• Comfort working with large, unstructured media datasets and the storage and compute tradeoffs they introduce.