Adopt Ssn Field with airSlate SignNow

Adopt ssn field

hello everyone my name is miriam mclemore and i'm an enterprise strategist with aws prior to joining aws i was the corporate cio at the coca-cola company i am thrilled to be your host for our new executive spotlight program where we will have the opportunity to have a deep dive conversation with aws leaders about business challenges our customers are facing we will also be able to address questions that leaders like you have submitted and today i have the the wonderful opportunity to have swami who is the vice president of amazon's machine learning with us today swami could you introduce yourself and tell us a little bit of your background sure first of all really excited to be here talking to you miriam um just in the way of intro so my name is swami siva subramanian i'm the vice president in charge of machine learning in aws and amazon so my organization uh works on all areas around machine learning with the goal to make machine learning accessible in the hands of everyday developer and data scientists and as uh personal uh nerd i've been with aws and amazon for 15 years since the early days of aws uh so really excited to be here leadership is such a challenge um just normally in in the speed of change that is facing all industries but in these times even more so are there key themes that you're hearing from customers that you're speaking to that have emerged in the last six to eight months yeah actually one of the biggest things i would probably tell is that given the unprecedented times we are living through now more than ever the uh advancements in technology matter more there is an increasing uh number of organizations who are forced to actually move to digital uh transformation much quicker so gone are the days where cios and ceos used to think digital transformation is a long term path and i can actually take a multi-year strategy when this pandemic started and a lot of actually worked from home started and various folks needed access to information suddenly digital transformation uh stopped being a multi-year project it needed to happen right now and that is one of the scenarios why we started seeing so many people started moving to cloud faster and adopting ai and ml technologies also much faster and i would generally put these things in four buckets uh if you were to think about it number one is many organizations uh right after we all started working from home and actually governments were to uh uh were asked to force to interact with customers digitally and so forth the number one challenge they hate is scaling customer communications and then uh the second one i would probably think about as the second bucket is around increasing agility in the organization suddenly the rate of change around the environmental respiratory whether you're suddenly i have to change your supply chain with the change in supply and demand or you have to think about uh how to alter uh your change changing your customer patterns or you have to process a new kind of workload where you can't wait for three weeks approval in processing alone and it needs to happen over the weekend and so forth so you needed to change and increase your agility immensely and then of course uh this time we also needed to uh every organization wanted to figure out how to get employees to work safely and what does it mean especially for folks who had to have a physical presence in the front line and finally also enable health care researchers and former companies on doing things like how to accelerate treatment how to do drug discovery faster how to exchange information faster so i generally see these major four buckets where organizations are moving very very fast and so let let's dive into those a little bit swami you talked about for scaling customer communications can you give us some examples of things that you've seen from customers you work with sure as here i mean as we all actually started working from home and then also as the pandemic started one of the number one things uh suddenly uh started happening is uh people were very uh uh eager to get more information from their public sector organizations it can be like the local government from the city or from the state or from the country as well and uh so there they needed uh increasingly quick and reliable way for them to get the latest on saying like where is the closest testing center or um what is uh when are the testing centers ours are open and so forth so this is where when you look at it one example that comes to mind is gaucher it's south africa's largest citizen engagement platform that is connecting over 50 million citizens onto more than 10 000 public representatives in the government so this connection between the citizens and the government is such a critical thing and one of the challenges is how do you scale this kind of massive number of citizens on communication to this small number of representatives and get them access to the current information so this is where gaucho actually leveraged one of our artificial intelligence service called amazon lex which enables developers to build a conversational chatbot interface quickly and they can deploy it live on the web or inside their contact center so couch at le leverage amazon lex to build an uh chat bot that can help citizens find things like uh what is the closest cover 19 testing facility and also it kind of enables citizens to stay more informed as well so on this chatbot i would speak uh we're able to exchange something like 14.2 million messages with no latency or backlog so think about the scale of what is the communication that a government organization was able to achieve again using artificial intelligence primitives like amazon lags now moving from there again it's not just about the government sector but also even in the private sector if you see organizations of all kinds and all sciences they were all it's really important for them to be able to scale their communications especially as more people were actually starting to engage digitally so swami those those were great customer examples can you talk a little bit about the importance also of increasing agility and efficiency in these challenging times the first one that comes to mind is a highly innovative company called cabbage so they are an online fintech company that are in the process of actually providing cash flow solutions for small businesses and cabbage has been using machine learning to quickly deliver uh things like financial relief packages for small businesses and they have been leveraging our ai technology called amazon techstrat which automatically takes care of uh things like taking a pdf form as an input and being able to automatically extract what is in the form and actually tell here is the first name last name and here is their ssn id and here are the relevant fields without having to go through and build custom ocr technologies and a lot of human uh processing and one of the most challenging workloads that they had was around the paycheck protection program where when the panamic hit cabbage was tightened where they enabled more than 500 000 small businesses employees to access paycheck uh protection program funding in the u.s where even their banks couldn't accommodate them so that is the incredible part so by using amazon text run cabbage took a week-long process of taking these forms and instead of having to do them manually now they shortened it to under 24 hours making a dramatic impact on the speed of how quickly they were able to process these loans so that's an example of how uh in our really agile startup we're able to actually use machine learning technology to augment human capabilities to serve customers faster and moving from like fintech to agricultural technology so one of the uh interesting uh things that we are starting to see in this world is that the supply chain uh is uh are changing with change in supply patterns and change in demand patterns so and uh especially in these times uh we all especially need to avoid any disruption in supply chain especially for essential items such as food i love your examples and what we can do with with data today another area that you mentioned was getting people back to work safely certainly all companies are trying to figure out you know what does that look like what are their mechanisms for keeping their team safe and are there technologies you know that could help them are there some that that you can share that have worked well for customers uh sure uh one of the uh i'll just uh highlight couple of them because we have seen like various interesting technologies built using many of our aws technologies right from our customers and our isv partners and so forth let me just highlight a couple first one is internally within amazon uh fulfillment centers and other areas we built uh something called amazon's distant assist so it is a computer version based technology where uh using uh cameras uh they were able to kind of project an uh image which kind of measures uh if uh social distancing is appropriately followed in any given uh field of vision so we have these in displays and uh location which serves as a visual reminder for all of us to say hey am i actually maintaining enough distance from my next colleague and uh make sure that we all actually can uh have a safe work environment uh if we were uh hard to assemble in the same place and what not and uh another example that is outside amazon is from queensland australia so there is an isv partner of aws uh on big mate they are a computer vision analytics company who built an uh thermal image uh scanning uh based on machine learning this application is called termi which tackles these issues using thermal imaging that can be deployed in any location so and uh you can deploy these uh thermal image cameras with these computer vision analytics uh and it can scan 30 people a second at any given time and run like 8.3 scans per second and the back end is essentially in a computer vision model running on amazon sage maker and these images are being sent to these models so set in the cloud and they are able to generate these analytics quickly in a timely fashion so again if you look at the common theme here uh again uh these are like great examples of innovation where uh distance assist which we actually even made at open source and available in github are actually big made which have made these technologies available for others to leverage as well these are great examples of being able to use machine learning to help people get to work safely yeah the is interesting the level of innovation that a crisis can drive so swami we talked about kind of the areas that you're seeing lots of innovation for our customers are there things that they're doing differently approaches that they're taking to machine learning and ai that that you didn't expect um yeah actually so uh um one of the biggest i would probably see it as uh some of the changes in the approach to machine learning and ai especially during these times the number one uh is that we are seeing a lot more focus on pragmatism over uh like open experimentation and just playing around with technologies what i mean by that is that it doesn't mean that these companies are not implementing machine learning but rather that they are getting very specific about what problem that they are trying to solve for their customers and they are really now starting to work backwards from the customer experience and scenarios and picking the right business case uh in fact in various other talks uh i have always talked about finding the right use case for machine learning is always the most important especially in kickstarting a culture for machine learning in an organization but it's even more important now so we have seen many of our customers use this moment as an opportunity to accelerate machine learning projects that will be very impactful uh for like machine learning safety or digital transformation or improving the efficiency all the way from like processing uh text-based documents and improving their workflows and so forth and they are de-prioritizing some of the more experimental projects uh and uh what not then the other one i would probably uh see is that they are now actually getting really good at implementing these projects also in parts and faces so uh they are not trying to actually do like a two or three year uh project instead many of them are actually focusing on what can i get done in the next few weeks and the next few months before you start expanding to any difference so this is where the cios don't talk about now multi-year investment technologies and road maps instead they are really now getting a lot more agile i think that's great to hear that customers are taking a more pragmatic and iterative approach certainly and in the conversations i have with customers that that is something that we talk a lot about is just get started find a a real problem that adds value to your business and drive that home and then do another right till you build the muscle and you learn as an organization how to leverage these technologies to your best advantage so i think you know again the crisis maybe has has provided some level of focus uh for organizations are there things that our customers have that are misconceptions about machine learning it's a great question actually so uh i would probably put them in like a couple of uh okay it's and uh the number one uh i would probably do is i see many organizations and really well-intentioned cios and ceos they typically uh they're so excited about machine learning and they want to actually move forward and then they end up hiring uh really good data scientists and put them in a central team and say let's go actually build out a few projects in theory this could work well but the problem is they end up getting in a centralized team so disconnected from the business units and the customers that what they end up doing is building a few proof of concepts or demos but they never get integrated into the mainstream business so this is fair i highly encourage all executives to almost hire these data scientists and embed them into the business unit so that they are closer to the customers they understand what problems need to be solved what are the challenges faced by the engineers and how to move quickly this also gets a shared buy-in within the business units where they actually see these uh ml projects as part of their customer experience and their business all together as well and then the second one is if you machine learning does not happen without data data is the fuel for uh machine learning and uh to get started on machine learning projects the number one thing you need to do is get your data strategy straight and this is where i encourage almost every cio to kind of first get your data strategy in terms of setting up data lakes setting up your access and control and governance and data preparation etl annotation all these things set up well and in fact even when i talked to various uh ml scientists in amazon or among our customers i often hear that more than 50 to 70 percent of the time they spend on ml projects is spent on various uh kinds of data processing so in fact when we tend to hire these machine learning scientists to build these amazing algorithms what they end up doing is spending a lot of time actually doing these uh data pre-processing and data wrangling and annotation work so this is where i think focusing on this data strategy as well as such an important thing to get it right as you've built out this capability at amazon and aws certainly you've had to experience failure to to innovate at the the speed that you have any kind of personal reflections on learning your way through this and and building out these capabilities uh sure uh i mean internally we always joke around uh saying like if you don't uh fail once in a while let alone fail often you're not innovating fast enough or innovating uh and pushing the boundaries hard so i think uh to a large extent uh failure goes hand in hand with innovation so and uh this is something where uh vs leaders uh uh in it can be in a startup or it can be in a big organization you want to actually if you're trying to drive innovation as a core principle and a culture within your organization i think you want to actually uh encourage your engineers and leaders to embrace failures along the way and provide them a safe environment for them to fail so because we all actually take risks when we actually take these big projects which have a lot of risk associated with them but if they end up actually feeling like hey uh my career is on the line if i end up failing uh in this project then uh you won't actually set up a culture of risk-taking so with that uh then i would also say to some extent failure is also a function of whether you stop trying after you hit a brick wall because uh there is a culture where you want to be constantly learning and reflecting on what you are finding out as you explore and that's what innovation is all about so and that is same as true even in machine learning where you're constantly experimenting with data and algorithms and seeing what works and uh then constantly iterating based on what you see and that is how you actually build an uh culture and machine learning and experimentation which is actually uh true even in innovation where you see our 10 that uh startups and even in amazon we actually almost every service team within aws operates like a startup uh its own startup if we actually we constantly work directly with customers and see what works and what needs to be uh what is not working so that we kind of actually constantly iterate on our approach based on customer feedback and pivot and uh this actually encourages us to constantly be innovating pushing the boundary and if we kind of uh hit any kind of wall then we know what to do to pivot and actually continue the innovation cycle as well absolutely and you know swami you've now been with amazon for i think you said 15 years and so at the beginning you know some people would think well you know it was a startup and you could fail but fail small as as the technology has changed and the spotlight's gotten brighter and uh you know failing and i i see this in large enterprises it it feels like you know you're going to fail big when you get into these big companies have you been able to to keep your edge and as you said keep it day one and still allow your team to fail small it's a great uh question and this is something we constantly uh think about uh as well so just to give even my own experience uh i think one thing that is kind of unique about amazon uh even 15 years ago i joined amazon as an intern so and uh one of the uh first things i remember in now one of the meetings early on was uh we were reviewing some uh database related issue that kind of caused disruption to amazon.com retail customers and uh i remember asking my manager uh who is who's werner vogel so because uh cto of amazon saying like why are we relying on this uh uh old relational database technology why can't we build something to the scale that amazon.com needs that provides some better durability and accuracy and scalability that he needs and and here is how we can do it and normally most companies would actually say you know what you're an intern you should go do what we the task we assigned to do but amazon is a little bit different so we actually encourage builders to build and we innovate so warner said it's a great idea if you have precise concrete thoughts write it down and go collaborate with this team which is thinking about this problem and that's how dynamo was born so and that's how now we made dynamo as an externally available database in the cloud called dynamodb which is now powering majority of amazon.com let alone actually huge number of thousands or tens of thousands of companies around the world as well so and that is like one example of also how amazon actually is again staying day one where if you unpack what happened here we actually are very very customer obsessed and seeing how can we continue to improve customer experience and we don't actually set up a culture of being top-down we encourage builders we want builders to build and we don't actually we encourage uh every builder to come up with innovative ideas and even senior leaders they are trained to say yes more than no and it's an important thing because ideas always don't come from the top in fact innovation always is coming from people who are closer to the customer and that is really the second element if you ask me one of the unique things i think about special about amazon is our leadership principles which are 14 of them because you can go check them out online but my favorite leadership principle is our customer obsession so if you see any meet again amazon when we talk about our product strategy our technical strategy and also how to evaluate competitive uh situation in the market or so forth we don't worry about competition or actually figuring out uh all the other things the number one thing we always think about is customers and we obsess over what is their experience how can deli how can we delight them and we work hard every single day to make sure that we are doing everything and building the right technology for the customer and this is one of the reasons now even internally within amazon we actually have like these single threaded teams uh for instance uh the uh sagemaker team operates like its own uh startup focused on actually helping ml builders build uh train and deploy ml models in the cloud and at the edge and that's all they are focused on and they are actively working with the relevant customers so so the decision making everything is actually highly uh close to where the customers are and the gm's act like their own ceos and this kind of culture helps amazon constantly actually make decisions faster and stay focused on the customers as well so and uh the other aspects of around enabling builders to build and actually come up with brilliant ideas is the other i would call it a secret sauce as well so swami what a great thought to end our conversation on customer obsession and putting the customer first thank you so much for your time and your insights it was terrific to have a few moments with you oh my pleasure again really enjoyed uh our conversation and uh very very uh excited as well to talk to you