AWS Floor28 News - AI/ML Special Edition

[Music] hello everybody and welcome to another special edition of AWS flow 28 news this time I would here with me Chris hi Chris yeah hi I don't good chris is part of our business development team for machine learning and AI and Chris tell me what do you do yeah I actually specific expensive specialized on rich media use cases okay and how to apply machine learning on video audio content and like get the essence of the video out of it let's go and you're visiting Israel for the next week a busy week yeah I see you got a very busy week couple of meetings multiple customers and we also have internal trainings but then also a meet up at the end of the week - yeah so let's speak about about let's talk about about machine learning and AI services what can I do as a developer with those kinds of services well the key point really is that you don't need to be a machine learning expert you can call api's and you will get the result based on the machine learning technology underneath but for example if you want to use like computer vision if you want to understand what's going on within your video what objects are in there what celebrities are in there you just need to call an API run the videos through it and you will get these results so basically any developer can use those API is with just few lines of code that's only an hour - of reading the commutation yes yes so that's that's the point that's for the developer obviously the stack is much deeper so it also serves anybody from the developer all the way down to the you know data scientist petitioner and so on but it really should support anybody so you mentioned video services like recognition what other types of services are we covering yes there's language services I'm not talking about the top layer of the stack so there's language services so like speech to text text to speech there is lacs which is a internal chat board and so forth but we also have like kind of very interesting services like personalized that a lot of media was using to create presentation engines or forecasts which timeseriesforecasting so these are like there's new services that were launched last year we could read man called guru code Cory Kendra four four four Enterprise Search so there is a very rich universe of services already well you don't need any machine learning so I assumed a different type of customers using this technology can you share with us customers interesting stories interesting implementations of those yeah so I'm obviously since I'm dealing a lot this video and audio so my customers more come from the media side the telecom advertisers gaming and so forth but they're across everywhere we use video right but like one example top of mind is what we did last year together with Sky News where you know we used recognition and these services to understand what's in the video of it was during the Royal Wedding understand what's what's in the video who is it arriving at the chapel and you know identifying that and then displaying a little blurb on it and like making that clickable and like making the experience you always ask yourself who is this guy who is it coming from right a little background story there and when did in the VOD version like when did he arrive and so forth and that was used with initial video correct okay so you're going to spend some time here in Israel what can you tell you tell me about the local market what Israeli customers or people well that it's also like we obviously talk about those specific use cases a lot of conversations around archive but also just content library understanding like tagging so what content do I have what's within my content and then build use cases on top of that right so once I understand what's going on there if there I don't know is the inappropriate content there is there do I look for certain objects and so forth I build a use case like I don't know a redaction use case or there's all discussions about subtitling and you know once I understand what's being said within the video actually the subtitling solution slave it automatically laid it automatically and things like that so it's really deep understanding in the content and then building use cases on top of that understanding and as far as I understand a lot of use cases that you extracting metadata from the video and then building some index int and being able to search on this search that that's the classical use case yes make it searchable discoverable for the editors but also for the end user I wonder what can customers do with this kind of notion after they getting this information from a video for example yeah well for example take you if your news organization right and you you have there was like an event and you wanna search in your archive if that specific person like what what content do I have about that very specific person specifically variant location or exactly and then it's much easier to discover well then you ask and archivists running down the corridor and it's like trying to find the tape right yeah and I can probably think about some cases for advertising as well yes correct so we work as a partner for example with mob they analyze in content and really have a deeper create a deep understanding of the content frame by frame and then they put it out there and they measure impact and then down to a deep third level like act on that impact right so this scene really resonates well with customers it is not that much so modify the content according and maybe little stage get customer advertisement based on exactly that's another thing right it has your context we're advertising but also for brand safety right so you don't want to have your brand associated with a certain event or whatever so there's a lot of there's hundreds of use cases or customers using these technologies yeah very interesting so I assume we've been traveling a lot are you coming from San Francisco why do you think about the local Israeli community or maturity in terms of machine learning what do you see in your visit well first it's super excite feels a little bit like San Francisco there's a lot of start-up there's a lot of vibe going on here so I really love it and then it's very advanced I mean I've been very very little time explaining machine learning or you know in general but you can do with this technology they go straight into the models they go straight into like how can I use it how can I bring it to customer quick so it's it's very advanced the conversation it's definitely one of the leading countries in machine learning worldwide okay I really hope you're gonna have a very interesting week yeah with your meetings and all other plans yeah yeah thank you very much for coming thank you for having me and thank you very much for watching AWS flow 28 new special edition about machine learning see you next time bye [Music]