Sachin Farfade

I am currently leading a science team in Amazon, India to develop Responsible AI (RAI) solution for Alexa+. The team is responsible for building end-to-end ML solutions for RAI which includes LLM model alignment, prompt tuning, building guardrail models and hotFix. Before moving to Alexa org I was working as a Principal Applied Scientist with International Machine Learning team at Amazon, India. Where I am leading science effort towards building ML solutions to provide navigational assistance to customers using Large Language Models. My effort include providing navigational assistance to customers to help in their shopping missions through customer experiences like use-case based shopping, refinement recommendation and AutoComplete. I have more than 20 years of industry experience working in conversational assistance, e-commerce, internet and embedded domains. Before joining Amazon, I worked at Yahoo Labs (Now Yahoo Research) with 3.5 years (2011-2014) stint at Bangalore and 3 years (2014-2017) stint at Sunnyvale, where I worked on developing computer vision solutions like face detection, face recognition and large-scale image tagging solution for Flickr using deep learning. Prior to that I worked with EADS-IW where I developed Content Based Image Retrieval (CBIR) system. In my first job, I worked for a start-up AllGo Embedded Systems where I developed hardware optimised video codecs and founded a Video Content Analysis team which is now spun off as a separate company called AllGoVision. I received my ME degree from Indian Institute of Science, Bangalore in Systems Science and Automation in 2005 and B.E. degree in electronics engineering from RKNEC, Nagpur in 2002. My research interest includes Computer Vision, Natural Language Processing and more recently Large Language Models. I have published in prestigious conferences like ICCV, WWW, WSDM and SIGIR etc. and re