Mohan Mahadevan, VP of Research, Onfido & Tony Fish, Founder, AMF Ventures
The interplay of Humans and Machines in AI-based Automation
Machine learning systems are now routinely performing complex tasks at unparalleled levels of performance across a wide range of applications. However, mission-critical applications such as those in financial services have a minimum tolerance for errors and error correction. Resultantly, the design of machine learning-based systems for such applications requires unprecedented levels of oversight and adaptability.
An optimal system has the right level of interplay between humans and machines. Onfido employs a large number of machine learning models to deliver scalable, secure and frictionless identity verification for their clients, whilst giving clients accurate data they need to remain KYC compliant—so ensuring their models deliver real-world results is business-critical. In this interview we talk with Mohan Mahadevan, VP of Research at Onfido, to learn about how the constraints on these applications, the tradeoffs in an optimal system, and what the future looks like.
Mohan Mahadevan, VP of Research, Onfido
Tony Fish, Founder, AMF Ventures
Mohan is an expert in computer vision, machine learning, AI, data and model interpretability, previously leading research efforts at Amazon as Head of Computer Vision and Machine Learning for Robotic Applications. He has over 15 patents in areas spanning optical architectures, algorithms, system design, automation, robotics and packaging technologies. As Onfido’s VP of Research he leads Onfido’s team of specialist machine learning engineers and is focused on ensuring their systems work both in the lab and the real world.