Job Details

Research Scientist/ML Developer

We are looking for a talented, experienced and highly-motivated machine learning developer/researcher with expertise in natural language processing, speech recognition and/or computer vision to join our engineering team in one of our Lisbon, Toronto or Silicon Valley offices.

ML developers at Talka lead efforts in developing cutting edge algorithmic solutions towards the end of improving communication between people. Your goal is to devise solutions that serve the short term needs of our clients but also conduct research towards the goal of better understanding among humans. You will work closely with our frontend and backend team to deploy your models into production. While currently, our company does not conduct research for academic purposes, such as submitting papers to peer-reviewed conferences or journals, our goal is to have an applied research department within 2 years.

Responsibilities

• Build and improve our current models on the modalities of text, audio and video;
• Interpret, synthesize, and analyze data using scientific or statistical techniques;
• Work with a team of annotators to improve their efficiency;
• Work with a team of frontend and backend engineers to deploy our models;
• Identify and understand business problems and propose corresponding ML solutions;
• Quickly analyze and come up with solutions to ML related problems;
• Mentor ML interns and junior researchers.

Requirements

• A master or Ph.D. degree in computer science or relevant area;
• Solid understanding of machine learning;
• Solid understanding of one or more of the following areas: natural language processing; speech recognition/audio processing; computer vision; information retrieval; and active learning;
• A solid publication record in top machine learning conferences or journals, and/or other measures of research output (e.g. github projects);
• Excellent Python skills;
• Experience with PyTorch or Tensorflow;
• Great communication skills;
• Proven ability to deliver quality code within budget and time constraints.