— Master’s degree in Computer Science, Engineering, or a related field
— 4+ years of professional experience working in Machine Learning, Computer Vision, NLP, Big Data, or a related field.
— Experience in at least one deep-learning framework (Tensorflow, PyTorch, etc.)
— Experience in at least one of the following: Computer Vision, ML for audio, natural language processing, ML model optimization
— Strong verbal and written communication skills
— Strong background in data and analytics
— Experience working with video—video capture, video processing, transcoding, frame analysis, ffmpeg
— Experience with distributed computation frameworks (Spark, Hive, Dask, Metaflow, etc.) and job orchestration (Airflow, Luigi, etc.)
— Conduct machine learning research on all aspects of model development to achieve state-of-the-art performance in terms of both accuracy and efficiency
— Design and own the curated, labelled datasets that are used for training models and evaluating them
— Define and implement the metrics used for the evaluation of models and the processes for model acceptance and QA
Our partner’s Core AI team is in charge of building models for the next generation of AI-powered TV products. They are responsible for the end-to-end development of their models, including
— dataset collection using geographically distributed television labs;
— model training in the cloud using serverless GPU clusters;
— model optimization for constrained computation on the edge;
— model testing using both virtual and real televisions; and
— creation of the data pipelines and tooling that makes the above possible.
As a member of their team, you will be working at the intersection of engineering, science, and entertainment. As a ML Engineer, you will work on the end-to-end development of our AI models. As an ML practitioner, you value efficiency, scalability, and repeatability.
Although these models lean heavily on the visual aspects of a video feed, they may also use the audio, subtitles, and related metadata; thus, computer vision, ML for audio, and natural language processing all come into play.