Next Day 2

Day two of NEXT was another day full of interesting interviews! Melanie and Mark sat down for quick chats with Haben Girma about accessibility in tech and Paresh Kharya to talk about NVIDIA. Next, we touched base with Amruta Gulanikar and Simon Zeltser to learn more about Windows SQL Server and .NET workloads on Google Cloud. The interviews wrap up with Henry Hsu & Isaac Wong of Holberton.

Haben Girma

The first Deafblind person to graduate from Harvard Law School, Haben Girma advocates for equal opportunities for people with disabilities. President Obama named her a White House Champion of Change. She received the Helen Keller Achievement Award, and a spot on Forbes 30 Under 30. Haben travels the world consulting and public speaking, teaching clients the benefits of fully accessible products and services. She’s a talented storyteller who helps people frame difference as an asset. She resisted society’s low expectations, choosing to create her own pioneering story. Haben is working on a book that will be published by Hachette in 2019.

Paresh Kharya

Paresh Kharya is Group Product Marketing Manager for data center products at NVIDIA responsible for product marketing of NVIDIA’s Tesla accelerated computing platform. Previously, Paresh held a variety of business roles in the high-tech industry, including group product manager at Adobe and business development manager at Tech Mahindra. Paresh has an MBA from the Indian Institute of Management and a bachelors of computer science and engineering from the National Institute of Technology, India.

Amruta Gulanikar & Simon Zeltser

Prior to joining Google Amruta spent 5+ years as a PM in the Office division at Microsoft working on many different products. Just before she left, she worked on launching a new service and supporting apps - “O365 Planner” which offers people a simple and visual way to organize teamwork. At Google, Amruta owns Windows on GCE which includes support for premium OS & Microsoft Server product images, platform improvements to support Windows workloads on GCE.

Simon Zeltser is a Developer Programs Engineer at Google, working with .NET and Windows on Google Cloud Platform.

Henry Hsu & Isaac Wong

Henry Hsu is a software engineer trained at Holberton School. He has experience with C, C++, Python, Ruby/Rails, JavaScript, HTML/CSS, MySQL/Postgres, Unity, Game Maker Studio, Linux, Photoshop, 3D Studio Max, systems design, algorithms, and devops.

Isaac Wong attends the Holberton School. He has a degree in horticulture from Texas A&M.

  • Edge TPU site
  • Cloud IoT Edge site
  • Cloud Armor site
  • Titan Security Key site
  • Building on our cloud security leadership to help keep businesses protected blog
  • Google Cloud Container Registry site
  • Haben Girma’s website site
  • Haben Girma’s presentation at NEXT video
  • San Francisco Lighthouse for the Blind site
  • National Federation of the Blind site
  • National Association of the Deaf site
  • NVIDIA site
  • NVIDIA and Google Cloud Platform site
  • Google Cloud Platform Podcast Episode 119 podcast
  • Velostrata site
  • GKE site
  • Google App Engine site
  • Stackdriver Debugger site
  • Windows on Google Cloud Platform site
  • SQL Server on Google Cloud Platform site
  • .NET on Google Cloud Platform site
  • Holberton School site
  • Unity site
  • GKE On-Prem site
  • TensorFlow site
Where can you find us next?

We’ll both be at Cloud NEXT in Moscone West on the first floor, so come by and say hi! We have chocolate!

[ARPEGGIATED SYNTHESIZER THEME PLAYING]MARK: Hi. And welcome to Episode Number 138 of our currently daily Google Cloud Platform podcast. My name is Mark Mandel. I'm here with my colleague, as always, Melanie Warrick.

MELANIE: Day 2, Day 2 of Next.

MARK: Day 2 of Next, indeed. It's been pretty great.

MELANIE: Yeah, it has been. It's been a busy day, lots going on here.

MARK: Lots and lots going on. So we've got a bunch of interviews happening today with people here on the show floor at Next. But what exciting things have you been particularly excited by in the keynote today?

MELANIE: The rotating stage platforms-- that's what I've been excited about.

MARK: Rotating containers, everything comes in containers.

MELANIE: Everything is in containers.

MARK: Exactly.

MELANIE: Some of the announcements are around edge computing, and specifically, edge TPUs and cloud IoT Edge, which I think is kind of exciting.

MARK: That's pretty cool. I really like Cloud Armor. I think that looks really great. I wish they'd spelled "Armor" with a U. But it's fine. I'll live with it. [LAUGHS] So that's a DDoS, an application defense system's services with GR based access control, which is really sweet.

MELANIE: Nice. And there's additional information out there about Titan Security Keys, a PHYTO security key with firmware developed by Google to verify its integrity.

MARK: I have one of the little Titan keys from last year. I don't know if-- yeah, yeah, yeah, the little ones. Also, there's a whole bunch of stuff Container registry vulnerability scanning, I'm particularly excited about that. So we have integrated stuff for when you have containers. We do security scanning so we can see if there's anything that shouldn't be fixed and stuff.

MELANIE: It's very important.

MARK: It is.

MELANIE: Mark, I think we got a bunch of people we're going to chat with today. So let's go do that.

MARK: Let's go do it.

MELANIE: So we are happy to have join us Haben Girma, who is an advocate for accessibility in tech. And I know, Haben, you're here with us at Next. You just gave a talk on accessibility in tech not too long ago. So thank you for being part of the podcast.

HABEN: Absolutely. It's an honor to be here. So my name is Haben Girma. And I am deaf/blind. I have limited vision and hearing. I can use technology when it's designed to be accessible. And it's my hope to teach more developers how to design products and services that are accessible.

MARK: So we have recorded a full interview with you which will be coming up soon.


MARK: But to tease a little of what we talked about previously, could you tell us a bit about what you talked about here Cloud Next?

HABEN: The main takeaway from my presentation this morning is that developers have the power to build services that everyone can access. A lot of people assume that people with disabilities don't use computers or don't use smartphones and ignore the disability community when building services. Don't make those assumptions. Design with everyone in mind.

There are 1.3 billion people with disabilities around the world. We are the largest minority group. People who invest in accessibility get to tap into this large market. So it's more revenue in the long run if you design with access in mind. So I want to encourage everyone to stop thinking about disability as charity. Instead, thinking of it as a business opportunity.

MELANIE: What's your favorite technology that's accessible and also just a technology everybody's using?

HABEN: I'm a huge fan of the iPhone because it has accessibility built in that lots of people with disabilities and non-disabled people can use.

MELANIE: So mobile phone usage in regards to what type of aspects work best for you?

HABEN: So with my phone, I now have easy access to communicate with people. Before that, I couldn't text. And without the ability to text, if I'm traveling around and need to talk to someone, get environmental information, maybe an emergency alert, I wouldn't be able to get that. But now, because I have a smartphone with accessibility built in, I have access to emergency alerts, to fun conversations, to restaurant reviews. And I have so much more freedom now with an accessible phone than I did before accessible phones were available.

MELANIE: What are some resources and outside organizations that you would recommend for people to check out if they're interested in understanding how to be better in designing accessibility in tech?

HABEN: My website has information about accessibility. So you're welcome to check out my website,, some local organizations, San Francisco Lighthouse for the Blind. You can also check out the National Federation of the Blind,, or the National Association of the Deaf. There are lots of organizations that can help provide information on accessibility.

MELANIE: What's been your favorite thing at Next, so far?

HABEN: I love meeting people. I've had the opportunity to meet lots of cool people here.

MARK: Excellent. Well, Haben, thank you so much for joining us on the podcast today. And look forward to listening to the full episode that will come out later in the next few months.

HABEN: Thanks for having me here.

MELANIE: So we're excited to have join us Paresh Kharya, great. He is from NVIDIA. And thank you for coming to join us today.

PARESH: Thanks a lot for having me.

MELANIE: So tell us a little about who you are and what you do over it at NVIDIA.

PARESH: Sure. So I am Paresh Kharya, responsible for product marketing for our data center products.


MARK: Fantastic. And I know you're here today-- NVIDIA is here today. There's a big NVIDIA presence. Can you tell us about what that's all about?

PARESH: Yeah, sure. So my company NVIDIA invented the GPU two decades ago that basically revolutionized the modern computer graphics and also revolutionized parallel computing. More recently, deep learning powered by the GPUs, have been responsible for advancements in artificial intelligence. And we are very happy about that.

GPU computing is also available across all the different computing form factors, from desktops to devices to servers to cloud services from our partners like Google.

MELANIE: Your here at Next because you had a big announcement yesterday. Can you tell us a little bit about that?

PARESH: Yeah, sure. We are very excited about that. So we announced that our Tesla P4 GPUs are coming to Google Cloud Platform. P4 is basically a small form factor, highly specialized for AI inference workloads GPU. And while our flagship GPUs like V100 have been available for a while, Google is the first cloud platform to make P4 available on a global scale. And this speaks to the tipping point of how AI inference is becoming really important as a workload.

P4 it is ideally suited for the real time inference workloads that require really fast processing. So think of services like speech-based interfaces or real-time live translations, understanding of video, and so on. P4 can process AI workloads that are really complex 40 times faster than what's possible on CPUs alone.

MELANIE: That's fast. Well, you probably heard us talk about this in the past in regards to inference and training. But first, do you mind explaining just briefly what is inference.

PARESH: Yeah, that's a great question. So training a neural network is when you're basically pumping in a lot of data to help a neural network learn by taking a look at the data and the labels that are associated with data. So that's the training step. It's computationally very complex.

Inferences is when you're putting that trained neural network to work. So it looks at new data, and it makes predictions. So for example, if it's about recognizing speech, it's taking in a new spoken sentence. And it's able to recognize it. So that's the AI inference.

MARK: That's really great. We should also make note that we have Episode 119, where we talked to Brian Catanzaro from NVIDIA, talking about deep learning as well, so if people want to hear more about that kind of thing as well.

MELANIE: Agreed, definitely. Well, what have you enjoyed or seen, so far, that you've been excited about outside of the announcement?

PARESH: Yeah, absolutely. So one of the really exciting things for me personally is Google's pioneering use of artificial intelligence in adding magic to your services, like--

MELANIE: Ooh, magic.

HABEN: --exactly, whether it's the G Suite or optimizing the performance of your massive data centers. And Google is not just keeping its technology to itself. It's basically making it available to everyone, every industry, and enabling them to be transformed by AI.

Google and NVIDIA, we have a strong partnership. And we both understand the need for accelerated computing to help advance AI. And Google continues to invest in TPUs, for example. But at the same time, Google is also relying on GPU used to power its own internal AI, as well as they are making it available as a cloud service on the Google Cloud Platform. So every deep learning framework and the diversity of modern AI can be supported.

MELANIE: That's great. Well, we appreciate you coming by to chat with us on the podcast. Anything else you wanted to mention before we let you go?

PARESH: Yeah, it's a really exciting time. We are witnessing a new revolution in terms of artificial intelligence and what it can do to revolutionize the workflows and simplify our lives, as well as revolutionize industries. And we're just at the tipping point right now, where people are realizing all of this is possible. We're very happy to work with Google to make our GPU computing technology available through the Google Cloud Platform to help companies and industries transform themselves.

MARK: Fantastic. Well, thank you so much for coming to join us here on the Next showroom floor.

PARESH: You're welcome.

MELANIE: Thank you.

PARESH: I appreciate it.

MELANIE: So joining us are Amruta Gulanikar, and we have Simon Zeltser. And we're going to talk a little bit about what's being presented in regards to Windows SQL Server and .NET workloads on GCP. So thanks for joining us.

AMRUTA: Thank you.

SIMON: Thanks for having us.

AMRUTA: Glad to be here.

MARK: Excellent. Do you want to tell us a little about who you are and what you do here at Google?

SIMON: Sure. So I'm Simon Zeltser. I'm a developer programs engineer focusing on workloads of Windows and .NET running on GCP.

AMRUTA: Hi. I'm Amruta. I'm a product manager on Google Compute Engine. And I own making sure that Windows and SQL Server workloads work well on the Platform.

MELANIE: Great. Well, what are you all excited about in regards to next?

SIMON: Wow, so--

AMRUTA: Lots of stuff. [CHUCKLES]

SIMON: --there are so many things to be excited about.

MELANIE: Yeah? What are a couple in particular around the whole Windows SQL Server and .NET workloads?

AMRUTA: So first is we're really excited about the fact that we have an amazing amount of migration tooling support available for our lift-and-shift workloads. There are some really cool demos about Velostrata, a company that we acquired recently. And that also has a lot of support for Windows workloads, especially.

We're also really, really excited about all the support that we have for .NET containers, which are running on GKE and GAE. And there's a lot of buzz around GKE and a lot of support for running workloads in a modernized fashion on Google Cloud. So we're excited to share-- Simon's going to be demoing some stuff, going to show you how you can take your legacy applications into the new world. So we're really excited to show some of those capabilities.

MARK: What are you demoing, Simon?

SIMON: I am going to demo two things. First is how we can lift-and-shift legacy .NET application into Google Cloud as is. And then the second demo is going to be about modernization of that application, splitting the monolith into microservices and then running it in .NET containers from Linux on GKE.

And also, we have a preview of Stackdriver Debugger, which allows us to do snapshot debugging of our .NET workloads in production, which is really, really cool.

MELANIE: That is great. Well, so any places that people can go look, in regards to additional resources, for Windows SQL Server and .NET workloads?

AMRUTA: Yeah, absolutely. So we have two sessions that are going to happen on Day 1. There is Session number 1 at 3: 15. It is around Active Directory. And we have a customer, Capital One, who's going to come and showcase how they brought their workloads to GCP.

We also have a follow-up session at 4:35, where Simon is going to do the demo that he just talked about. And we're going to talk about all the platform features that you have on support for Windows SQL Server and .NET workloads.

We also have a booth where we're going to have .NET experts to come and talk to us. And we have a ton of information on our website. So go to,, .NET, and so on and so forth. And we have a page with a lot of resources and links out there.


MARK: Fantastic. Is there anything else we've missed that you want to make sure that people want to know about?

AMRUTA: Well, come talk to us. We'll be hanging out, and we're excited about all the new stuff to show you.

MELANIE: Great. Well, thank you both for joining.

SIMON: Thanks for having us.

AMRUTA: Thank you for having us.

MARK: Thanks. So we have two completed students from Holberton School, Henry [INAUDIBLE] and Isaac Wong, joining us today. Henry, how are you doing?

HENRY: Great, doing great. Thanks for having me.

MARK: Excellent. Isaac, how are you doing?

ISAAC: Doing wonderful.

MARK: Wonderful. So why don't you tell us a little bit about Holberton School and your experience there?

HENRY: Holberton School is a two-year college alternative training software engineers. It's pretty much project based. There's no instructors. And it's just pure learning and project based to try to emulate a more realistic approach to learning, as opposed to just giving you assignments and that you do it. So yeah.

MELANIE: Thanks.

ISAAC: Yeah, it also has a deferred tuition model. So we don't pay anything until we get jobs. Unless we have good jobs and they pay us, we pay the school back for training [INAUDIBLE].

MARK: Nice.

MELANIE: What are some of the things you're studying?

HENRY: So we start off primarily by doing low-level programming, learning old-school C. And then after that-- we do that for about three months-- and then we progress on to higher level languages like Python and JavaScript. And then we also do some DevOps training also to learn Linux and bash Ls scripting, and then make some full-stack web apps. So that's the broad overview, condensed.

ISAAC: Yeah. And so I'm actually in class right now, and yesterday I was working on a MySQL project to make a backend for an Airbnb clone. And so it's ramping towards to doing DevOps as a kind of work. But this is a very interesting conference for us to attend.

MARK: So speaking--


--we are here at Cloud Next. What have you seen? What has been particularly exciting to both of you?

HENRY: I come from a health care background. So when I saw the speech that they were giving yesterday about how they're leveraging cloud technology to improve patient treatment, that was very interesting to me, especially since I'm trying to see if I can still translate some of my domain knowledge in the health sector into technology work. So hearing about that, that was very interesting to me.

On a less serious note, I'm also kind of a gamer. So when I went up there and I saw what they were doing with this little dancing robot, and then they also were showcasing some of the games that's made in Unity, which apparently transitioned from AWS to 2, that was really awesome. So just yeah.

MELANIE: Mark knows a little bit about that.

MARK: A little smidge. [LAUGHTER]

MELANIE: And there's an activity later today that you're specifically supporting. You want to talk about that for a second?

MARK: Yeah, so I'm doing-- there's actually going to be a gaming meetup, So there's a Birds of a Feather conversation happening at 4: 00 PM on Moscone 2.


MARK: It'll be this afternoon. So if you're listening to the podcast, unfortunately, this is no news to you.


MELANIE: You've missed it.


But you should talk to the people who went.

MARK: Yeah.

MELANIE: Anyways, well, so I know when we were starting out the podcast, you both were sharing some interesting knowledge I think would be very useful for others who go to conferences.


MELANIE: What are some best practices on gathering swag?

ISAAC: OK, so if you're a shameless poor student like Henry and I, what you do when you go to a conference, you find a booth that has swag bags. You know those fancy bags?

MARK: Yep.

ISAAC: And then you use that bag to go around to other booths to get candy, chocolate pens, robust trinkets that you can show around from the conference. Yeah, so most people are very willing to give away candy. I think Mark and Melanie were offering us whole bowls of candy here.

MARK: Yeah.

ISAAC: As long as you ask for it, you're fine taking it. If you just go over there and start pouring the whole candy thing into your bag, yeah, it's not as nice. Ask, ask.

MARK: Isaac, what cool stuff have you found particularly great?

ISAAC: Yeah, so I might be naive, but I thought it was really interesting when they announced the GKE On-Prem server.

HENRY: Oh, yeah.

ISAAC: I thought that was really cool. I'm not too involved in that kind of work yet. But that seemed like a very nice innovation there. And I'm really interested to see how that develops.

MELANIE: And you're right. It is a very cool innovation.

ISAAC: And so me personally-- so Henry is from the health care background. I'm actually from the agricultural background. And so I think-- I'm not sure if it's today or tomorrow, but I'm interested in a session tomorrow by a company that's trying to incorporate machine learning in the field of agriculture to see how that [INAUDIBLE].

MARK: Excellent. Have you both been going to sessions while you're here as well?

HENRY: I plan to go to more of it today. After this, I'm going to make the rounds.

MARK: What sessions are you planning on going to?

HENRY: Well, I want to try to hit some of more of the-- again, the (LAUGHING) gaming and entertainment ones.

MARK: Cool. I'm OK with that. That's my jam. That's my thing, so that's great.


MELANIE: Was this planted?


MARK: No, it wasn't. It wasn't. Any particular gaming sessions you want to go to?

HENRY: I saw that they have the booth, the dancing robot one. I didn't actually get to see it in session. I was kind of waiting. But then I saw they were showing the dotless showcase. And then so I got to talk to Trey. And he was explaining to me. And I learned a lot about how containers, actually, in virtual machines are being used to run the matchmaking servers, and stuff like that in the background. So it was really interesting.

MELANIE: Wait, there's a dancing robot? How did I not know about this?


Why am I not there right now?

MARK: They have a really amazing exhibit up there where they've got an animator here. And I think the other animator is in LA.

ISAAC: Oh, wow.

MARK: And they're doing live animation stuff on the show floor on the entertainment side. It's really cool.

ISAAC: I need to go to a booth with you. I mean, I've been wanting to go to Friday.


MARK: Awesome. Yeah, let's do it.

ISAAC: I'll come with you. Yeah, so I went to a TensorFlow session earlier today. It sounds like I'm really interested in machine learning. I just don't know anything about it. So I'm here to learn about it. So that was real interesting for me.

MELANIE: Good reason to be here.

ISAAC: Yeah, I mean, I was intimidated because going through the whole session list, there's a whole bunch of things which I don't know about. But it's nice that they have introductory sessions like that for people like me to go to.

I went to the Hands-On Lab too, again, another good introductory session too for people who don't know anything about the Google products and it's machine learning, TensorFlow, things like that. So that was cool.

MARK: Fantastic.

MELANIE: Everybody has to start somewhere. So that's great that you're doing that.

MARK: Yeah.

MELANIE: Well, thanks, guys. We really appreciate you coming in to share with us your experiences about Next.

HENRY: Thank you.

ISAAC: Thank you for having us.

MARK: Thanks.

MELANIE: Thanks, everybody, for joining us again today on Day 2 of Next. So Mark, I guess we got one more day left, huh?

MARK: One more day left. But I think it's going to be a really good day tomorrow.

MELANIE: It'll be the developer day tomorrow.

MARK: Yeah, so expect cool things to happen.

MELANIE: And if you're here, come by and say hi. We're on Moscone West, on the first floor. And we have lots of chocolate. We have podcast branded chocolates.

MARK: We do.

MELANIE: You should come grab some.

MARK: Absolutely. You definitely should.

MELANIE: Have a good one.

MARK: Thanks, and have a great Next.



Mark Mandel and Melanie Warrick

Continue the conversation

Leave us a comment on Reddit