Research Engineer

Sept. 11, 2020

Remote, USA

$90,000 - $120,000 / yr

We're looking for a Research Engineer to help expand our multimodal information processing systems for video understanding.

Machine learning for video understanding requires multimodal information. As a research engineer, you'll have the opportunity to work on projects spanning multiple machine learning topic areas, including computer vision, natural language processing, and structured knowledge graphs.

Typical activities include implementing and evaluating deep learning models, integrating features from multiple video modalities, curating datasets for novel machine learning applications, and deploying models into production. Example projects: multilingual optical character recognition, webly-supervised object detection, language modeling for speech-to-text, and knowledge graph representation learning.

Typical tasks

  • Discover and evaluate relevant models and datasets for machine learning applications
  • Design and implement high-capacity machine learning and data analysis pipelines
  • Partner with product and business development teams to align system design and behavior with usage expectations
  • Deploy pipelines on cloud infrastructure

Essential background

  • Demonstrated experience implementing and training deep neural networks in Python
  • Applied research experience in at least one of the following domains: computer vision, natural language processing, knowledge graphs
  • Great written communication and documentation skills
  • Familiarity with object oriented design practices
  • Skill with version control, package management, shell, etc, is assumed
  • At least one of the following
  • PhD in computer science, applied mathematics, or quantitative sciences
  • Master's in the same with 1+ year of industry experience with applied deep learning
  • 3+ years of relevant industry experience with applied deep learning

Even better

  • Publications in applied or theoretical machine learning
  • Open-source applied ML projects or contributions to ML libraries
  • Plenty of practical experience with deep learning using PyTorch

About Vidrovr
We develop an end-to-end video search and understanding platform that indexes and monitors live streams, periodic video feeds, and manually uploaded video content. We work with private, public, and non-profit partners to provide solutions for a variety of large-scale video meta-analysis problems. Public applications powered by our platform include CNBC's Warren Buffet Archive, Alliance for Securing Democracy's Hamilton Dashboard for monitoring state-sponsored disinformation, and shot-list generation for the Associated Press's video footage archive.

Built with ❤️ for the ML Community by Dom © 2020 RemoteML