Shirsha currently works in Thomson Reuters Labs leading the Engineering team in Bangalore. In her current work she is responsible for successful delivery of AI ML projects for internal platforms, tools and products. Having experienced the ML lifecycle and the tangent that it has to the software development lifecycle, she will share on machine learning operations and what it takes to land a solution with ML!
In her prior avatar she has worked in Mercedes Benz R&D India, Nokia and other companies, on building and deploying solutions with Big Data, Analytics and ML.
She started her career as a software engineer building solutions for handsets in Motorola, moving onto 3G call processing software, network probes. It was a gig on data processing for location based services that immersed her in the world of Big Data, Analytics, pivoting her career to the ML space now.
She holds her experience as a software engineer special and hopes to see the level of standardisation in ML projects as exists in regular software projects.
The need for Testing in Machine Learning Projects
The world is crazed by the popularity, possibilities and wonders of a solution that naturally embeds AI. Tech companies, big and small, are racing in their journey to embed AI into their existing solutions and/or create new ones. In this talk, we reflect on impacts that legacy software solutions, workflows, business processes and large chained tech systems have with the introduction of AI-powering services into them. We lay emphasis on the various needs for testing that introduction of ML for a large software solution. We hope to highlight some of the requirements we have uncovered as part of our Machine Learning projects for testing. We conclude with feedback and requirement analysis from planning POV for AI solutions as part of this talk.