Intel’s Robert Blum joins host Jake Smith to talk about Intel’s broader connectivity group and how the company facilitates efficiently moving data between servers and across the data center with a wide range of products, including silicon photonics, Ethernet switches, accelerators and more. Robert discusses the advancements Intel has made in silicon photonics over the last four years since first launching 100GB transceivers and highlights the milestone of having shipped more than 5 million transceivers for hyperscale data centers to date. He goes into more detail about the industry’s explosive growth and the need for 800GB transceivers, which are now being sampled by customers in the field. Jake and Robert talk about the benefits of co-packaged optics with Barefoot switches, which are evolving from proof-of-concept to deployment in the coming years. Robert also touches on the future of silicon photonics, such as Mobileye’s application for LiDAR. 

Learn more at: https://intel.com/siliconphotonics
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

The Future of Silicon Photonics at Scale - CitC Episode 230

sMedio’s George Tang joins host Jake Smith to talk about why the Japan-based, multi-media company looks to Intel to help with cost-effective solutions to train their facial recognition modules and cater to their scaling needs. The two discuss how Tensoflow optimizations helped improve training performance on Intel Xeon processors, the cost of cloud-based AI projects, the types of robotics products sMedio is developing, and the benefits of working with the Intel AI Builders program.  

Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Optimizing AI for Japanese Robots - CitC Episode 238

Accubits’s Shameer Thaha joins host Jake Smith to talk about machine vision and predictive analytics. Shameer explains the use cases for the company’s Emotyx platform, such as real-time crowd analytics, people detection, counting, tracking, emotion detection, health care and more. Shameer goes into detail about the optimizations of the Intel Distribution of OpenVINO, using a single CPU system without GPU support at the edge, and some of Accubits’s projects around the world.

Follow Shameer on Twitter at: https://twitter.com/shameerthaha
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Machine Vision Around the World - CitC Episode 242

Sharath Puranik, Director of Product & Engineering at Stylumia, talks with host Jake Smith about how Stylumia leverages machine learning to improve product assortments, optimize inventory management, and offer a consumer-driven forecast to reduce waste in the fashion and lifestyle industry. Sharath explains why Stylumia became a part of the Intel AI Builders programs, which provided the support that allowed the company to transfer the majority of their models off more expensive and costly GPUs to more scalable Intel CPU architecture with the OpenVINO framework. 

Follow Stylumia on Twitter at: https://twitter.com/stylumia
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Using AI to Forecast Fashion in India - CitC Episode 245

SigOpt cofounder and CEO Scott Clark joins host Jake Smith to discuss his company’s recent acquisition by Intel, how this acquisition accelerates their ability to meet their mission to empower AI engineers, and the latest innovation in the SigOpt intelligent experimentation platform that enables faster design, exploration, and optimization of models. Scott goes into detail about how SigOpt software helps developers track their modeling artifacts, design experiments, explore their parameter space, and manage hyper parameter optimization (HPO). Jake and Scott close the episode by talking about the future of AI. 

Learn more about SigOpt at: https://sigopt.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Experimenting with AI Optimizations - CitC Episode 250

Taboola’s Ariel Pisetzky joins host Jake Smith to talk about using artificial intelligence (AI) for personalized content recommendations at massive scale, reaching up to 4 billion web pages a day. Ariel discusses why Taboola runs their services on their own on-prem infrastructure of over 10,000 servers with Intel processors, many of which are dedicated to inferencing workloads. Ariel also goes into detail about some of the performance optimizations the company has found gen-over-gen with increased memory access. Jake and Ariel close the episode by talking about the open source community and how to respond to a crisis, like a fire breaking out inside your data center. 

Follow Taboola on Twitter at: https://twitter.com/taboola
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Behind the Massive Scale Required for Content Recommendation - CitC Episode 252

Soundar Rajan from Perpetuuiti Technosoft Services joins host Jake Smith to talk about software automation solutions for IT and business operations. Soundar goes into detail about Robotic Process Automation (RPA) with the company’s Av3ar platform—a human-like AI+RPA platform of intelligent BoTs (or iBoTs) which can absorb, deconstruct, and use information as a human would to resolve problems. He also explains how OpenVINO is helping the company create new solutions and increase customer performance.

More information available at: https://ptechnosoft.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

The Future of Robotic Process Automation - CitC Episode 257

Sanjay Prasad from Metaswitch (a Microsoft company) joins host Jake Smith to talk about why Metaswitch collaborates with Intel to improve network function virtualization, media transcoding, throughput and more. Sanjay talks about some of the performance gains his team has seen over the last ten years from gen-over-gen processor advancement and discusses the future communication service providers can expect, including the rise of microservices. 

Follow Metaswitch on Twitter at: https://twitter.com/metaswitch
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

The Future of Cloud Native Communications - CitC Episode 259

Nitin Somalaraju from Tech Mahindra joins host Jake Smith to talk about the company’s Speech Analytics platform named Sayint, which uses cutting-edge artificial intelligence (AI), machine learning (ML), and natural language processing (NLP) to uncover meaningful insights from customer conversations. Nitin talks about why regional localization and privacy issues are still challenging to overcome when deploying voicebots and how Intel technology has helped the Tech Mahindra team.

For more information, visit: https://www.sayint.ai/

Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Leveraging Conversational AI to Improve Customer Experience - CitC Episode 263

Arijit Sengupta, founder and CEO of Aible, joins host Jake Smith to talk about the difference between the promise of AI and actual results. Arijit talks about helping develop BeyondCore, which was acquired by Salesforce, his book AI is a Waste of Money, the current unsustainable costs of creating complex models, and why Aible takes a “serverless-first” approach to AI. He also goes into detail about working with Intel on optimizations and benchmarking for necessary lower energy costs. Jake and Arijit close the episode discussing algorithm fairness, data privacy, and the personalized AI of the future.

Follow Arijit on Twitter at: https://twitter.com/arijit_sg
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

The Future of Serverless AI - CitC Episode 264

Erik Landolsi, CTO of Irisity, joins host Jake Smith to talk about using AI for video analytics. Erik talks about the challenges of connecting and processing immense amounts of video data, why Irisity anonymizes certain details based on where cameras are installed, and what the recent acquisition of AgentVi means for the future of specialized video analytics. 

For more information, visit: https://irisity.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Using AI for Specialized Video Analytics - CitC Episode 267

Atif Kureishy, founder and CEO of Vistry, joins host Jake Smith to talk about deploying restaurant automation technology improve quality of service. Atif goes into detail about Vistry’s Discrn platform, why the company extended Intel’s edge insight for industrial software architecture to optimize workloads, and how his team uses the Intel Distribution of OpenVINO for faster insights and increased performance.

For more information visit: https://www.vistry.ai
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

AI Automation in the Restaurant Industry - CitC Episode 269

Markus Lippus from MindTitan joins host Jake Smith to talk about how AI is innovating many industries across the world including customer service. Markus discusses how his background in biology shaped his passion for machine learning (ML) and helped drive the creation of Mind Titan. He highlights how MindTitan and Intel worked together to optimize their TitanCS solution on Intel Extension for PyTorch achieving improvements on inference performance running on Intel CPUs and driving better real-time call analysis. Markus also talks about how MindTitan’s solutions are driving disruption in the telecommunications industry citing an important customer case study with Elisa Estonia AS. Jake and Markus end the episode talking about the future of AI research and development.
For more information, visit: https://mindtitan.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

AI-Based Customer Service Automation - CitC Episode 270

Anup Mehta, Co-Founder and CEO of DeepEdge, joins host Jake Smith to discuss DeepEdge’s vision to enable new edge applications, bringing the power of deep learning and computer vision to the hands of the users. He chats about how DeepEdge's deep learning operations platform supports the entire machine learning lifecycle from on boarding data to deploying edge optimized models to Intel hardware. Anup talks about how the Intel OpenVINO toolkit is seamlessly integrated with the platform for model conversion, optimization, and deployment. He also illustrates how the DeepEdge platform leverages other tools from the OpenVINO toolkit including post optimization toolkit and Neural Network Compression Framework. Anup also highlights how their platform can enable customers to migrate their workloads from a GPU based infrastructure to Intel architecture-based platforms at a fraction of the cost. Lastly, Jake and Anup dive into discussing the future of AI and how it can solve many every-day problems that people face today and how AI can deliver value to everyone in the world. 
For more information, visit: https://deepedge.ai/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

No-code Deep Learning Development with DeepEdge - CitC Episode 278

Vik Anantha, Chief Product and Strategic Marketing Officer at Caresyntax joins host Jake Smith to discuss how Caresyntax’s data-driven surgery platform is helping make surgeries smarter and safer. Through applying AI and analytics automation, he describes how the Caresyntax platform enables clinicians to focus only on necessary variations for a surgery and in turn optimizes the clinical, operational, and financial outcomes. Vik talks about how the Caresyntax platform can reduce unwarranted surgical variation from pre-operation to post-operation by assisting surgeons and interventionalists to make better decisions in real time utilizing important data captured during surgery. He also points out that the platform is optimized for the Intel architecture to bring performance and flexibility to their customers. Lastly, Jake and Vik talk about what the future holds for AI and how Caresyntax in collaboration with Intel is accelerating AI’s ability to solve real business problems in healthcare. 
For more information, visit: https://caresyntax.com/ 
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Data-driven surgery with Caresyntax - CitC Episode 279

Lorenzo Fornaciari, Vice President of Strategy iGenius joins host Jake Smith to discuss how iGenius is reimagining data analytics and bringing it to the next level with their platform named Crystal. Lorenzo describes how Crystal, auto-classifying business requests in multiple languages, enables businesses to derive real-time insights and access key insights from their data without any training or data literacy skills needed. He highlights how iGenius worked with Intel to optimize their natural language processing and discusses the future of business analytics through the idea of the augmented consumer, a consumer that is able to use new augmented analytics technologies to easily explore and derive insights from their data. He describes how augmented non-technical data consumers represent an un-tapped user base that owns many decision making process and rely on smart business intelligence platforms like iGenius. Lastly, Lorenzo discusses how iGenius is using AI to remove the barriers between the businessperson and data in order to democratize the insights that are hidden and available in the data to boost a decision-making process for every enterprise.  
For more information, visit: https://www.igenius.ai/ 
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Empower Data Analytics Business Users with AI - CitC Episode 281

Renata Morato, PO & PMO at HooBox, joins host Jake Smith to discuss how HooBox is utilizing AI and Machine Learning to better patient experiences in healthcare settings. She highlights how HooBox’s SADIA platform uses AI to improve many different aspects of hospital patient care including fall risk detection, pain levels, and pressure injuries. SADIA can address all these issues with only one camera mounted on a hospital bed. Renata describes how SADIA can be an advanced technology pathway to reduce length of stay and patient’s complications by democratizing AI enhanced patient care. Lastly, Jake and Renata discuss how HooBox is helping the healthcare industry move to a new standard providing closer assistance to patients, across all sorts of data, moving from reactive to predictive.

For more information, visit: https://hoobox.one/ 
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

HooBox Using AI to Improve Patient Experience - CitC Episode 283

Ankit Narayan Singh, Co-Founder & CTO at ParallelDots, joins host Jake Smith to discuss how artificial intelligence and image recognition are transforming the consumer-packaged goods (CPG) and retail industry operations in brick-and-mortar stores. He describes how it's a challenge for retailers to identify and replenish out-of-stock products on shelves quickly and efficiently. Ankit highlights how ParallelDots’ ShelfWatch platform analyzes photos captured in stores and uses computer vision models to send real time alerts to retailers allowing them to rapidly analyze and refill any shortages on their shelves. Ankit also illustrates how ParallelDots collaborated closely with Intel leveraging Intel’s ecosystem program to optimize their DL models for OpenVINO. This collaboration helped significantly increase their inference throughput performance and also enabled ParallellDots to optimize some of their models to run on CPUs avoiding the need to run on more expensive GPUs. Lastly, Jake and Ankit talk about the digitization of brick-and-mortar stores and how that process is building the stores of the future. Computer Vision and artificial intelligence are exposing more data and greater capabilities for retailers to provide better experiences to their customers while increasing efficiencies.

For more information, visit: https://www.paralleldots.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Generating Real-Time Shelf Insights with ParallelDots Computer Vision Platform - CitC Episode 284

Mouli Narayanan, Founder and CEO at Zeblok, join Jake Smith to discuss Zeblok’s Ai-MicroCloud platform and how it streamlines AI delivery by providing a single, cohesive, turnkey, cloud-native AI environment. Mouli shares Zeblok’s mission and describes how the Ai-MicroCloud enables enterprises to quickly deliver AI at the edge and integrate AI solutions into mission-critical processes, faster and with lower costs. He expresses how AI deployments at the edge are becoming a main point of data origination and consumption for the enterprise today. Discussing how Zeblok is helping drive this shift and assisting enterprises with needed ISV (independent software vendor) integration to drive their edge AI deployments. Mouli also covers how collaborating with the Intel Disruptor Program has enabled Zeblok to optimize their platform for Intel technologies like OpenVINO and bring greater value to their customers. Lastly, Jake and Mouli discuss the future of AI and how artificial intelligence will drive incredible improvements in productivity in the enterprise.
For more information, visit: https://www.zeblok.com/ 
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Cloud Native AI Development with Zeblok - CitC Episode 285

Yonatan Geifman, Co-Founder & CEO at Deci, joins host Jake Smith to talk about how Deci’s AutoNAC (Automated Neural Architecture Construction) engine enables developers to redesign their deep learning models to significantly improve latency performance while preserving the model's accuracy. Yonatan further explained that improving latency and accuracy would allow developers to shorten the development cycle and deploy the optimized model to production environments faster. Yonatan shared with Jake the close and deep partnership Deci has developed with Intel in engineering collaboration and ecosystem programs. When asked about his thought on the future of AI, Yonatan said while it would take more time for AI to be everywhere, AI is being democratized by startups, large companies, and open-source communities, so more and more people will be able to build AI applications in the future.  
For more information, visit: https://deci.ai/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Optimizing Model Performance with Deci - CitC Episode 286

Karl-Heinz Hoffmann, VP of Technology Ecosystem at SingleStore, joins host Jake Smith in a conversation about how SingleStore's database technology (SingleStoreDB) unifies transactions and analytics to power real-time business decision-making. Karl-Heinz talks about modern business applications where companies take advantage of SingleStoreDB to bring real-time analytics and data feeds together, driving business impacts. Furthermore, Karl-Heinz shares his view of the future database structure moving towards an architecture that consolidates transactions and analytics to accelerate real-time business insights.

For more information, visit: https://www.singlestore.com/
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

How SingleStore Powers Real-Time Analytics – CitC Episode 287

Joining Jake on the podcast today is Kasia Borowska, Co-Founder and Managing Director at Brainpool AI. The duo discuss the most popular applications of AI in business, how they benefit, and the best ways to start implementing solutions.

Brainpool AI is an Artificial Intelligence services company that connects a global network of over 600 AI and Machine Learning experts with corporate partners. Intelligent automation helps their clients grow efficiently, reducing operational costs and overhead. The resultant higher quality products and services greatly benefit their customers’ clients.

Libraries like Intel OpenVino help leverage existing models in creating solutions for business partners. Borowska finds most customers would like to avoid “reinventing the wheel” when it comes to onboarding new AI solutions.

Check out their website for more information at brainpool.ai . Their LinkedIn also offers updates and industry trends.

Going Deep with Brainpool AI and Intel OpenVINO - CitC Episode 288

Davit Buniatyan, Activeloop CEO, joins host Jake Smith on the latest episode of Conversations in the Cloud to discuss optimizing data lakes for creating deep learning-based products. Deep Lake allows storing complex data in AI-native format and streaming it over the network to Tensor Query Language, in-browser visualization engine, and deep learning frameworks without sacrificing GPU utilization. A member of the Intel Disruptor program, Activeloop partners with Intel to deliver Deep Lake on Intel® Xeon Scalable Processors and make Deep Lake more performant.
 
Learn more about Deep Lake at activeloop.ai  

Research Links: 
Deep Lake Whitepaper: https://www.deeplake.ai/whitepaper
Deep Lake Academic Paper: https://arxiv.org/pdf/2209.10785.pdf 
Deep Lake GitHub Repository: https://github.com/activeloopai/deeplake
Deep Lake Datasets: https://datasets.activeloop.ai/docs/ml/datasets/

Building a Data Lake for Deep Learning with Activeloop - Episode 289

Dr. Yichuan Zhang, CEO of Boltzbit, joins host Jake Smith to discuss what is generative AI, why it matters to developers, and how the Boltzbit Generative AutoML platform helps tackle enterprise AI challenges. Dr. Yichuan started by sharing his journey of transitioning from academia to industry and entrepreneurship while developing a strong passion for solving complex challenges with generative AI. Dr. Yichuan explained why the next phase of AI is generative in terms of bigger and more dynamic output capacity, which makes AI development easier and faster. Dr. Yichuan praised the partnership with Intel where software optimization on CPU makes AI accessible on a broader range of devices, especially in public cloud infrastructure. While recognizing AI still has a long way to go, Dr. Yichuan sees generative AI to play an important role in unleashing AI capabilities to catch up with human intelligence at some levels.

For more information, visit: https://boltzbit.com/ 
Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Boltzbit Generative AI Tackles Enterprise AI Challenges - Episode 290

On today’s episode of Conversations in the Cloud Alex Liu, Chief Technology Officer of BasicAI stops by to talk about the company’s efforts at evolving the training data platform. Xtreme1 is the world's first open-source platform for multisensory training data and can rapidly accelerate the processing and management of training data. 

AI engineers spend a preponderance of their time preparing the training data. Through advanced AI-powered annotation tools, Xtreme1 improves modeling. Capable of handling various time-sensitive online/offline tasks, distributed data analysis computation, model training, evaluation, and inference, Xtreme1 is fully compatible with CPU-only runtime environments. In particular, the model inference has been optimized for Intel Xeon CPU to provide higher throughput. Inference throughput on Intel Xeon 8380 is vastly improved with the help of Intel extension for PyTorch and OpenVINO toolkit.

For more information go to www.basic.ai and www.xtreme1.io

BasicAI Leads Multisensory Training Data Effort with Xtreme1 - Episode 291

Deepak MK, Vice President of Data Science at ExamRoom.AI sits in with Jake to discuss the company’s primary mission: to provide a highly secure, unified assessment system and fair environment to the candidates who take online remote proctored exams.

ExamRoom.AI’s research and advancements in AI technology have unleashed leap-forward technological accomplishments, leading to efforts toward making robust online testing a reality. Universities and organizations alike face difficulties in ensuring integrity and security in testing that ExamRoom.AI can address.

Using 3rd Generation Intel® Xeon Scalable Processors and OpenVINO helped ExamRoom.AI reduce the occurrence of memory issues and build better machine learning models through its scalability and latency reduction, greatly reducing model training time. Intel’s Ice Lake processor speeds up and enhances their systems, helping avoid dreaded Out Of Memory (OOM) errors.

Check out their website for more information at https://examroom.ai/

Computer Vision and NLP in Practice with ExamRoom.AI - Episode 292

Learn why our customers are moving to immersion cooling, and how it and sustainability are changing data centers in this conversation with Intel Fellows Mohan Kumar, Nishi Ahuja and Senior PE Sandeep Ahuja and host Jake Smith.

For more information, visit: Data Center Sustainability- Intel and Submer Partner on Immersion Cooling | Intel Business - YouTube https://www.youtube.com/watch?v=yBVO_YQBYjc  

and Intel® Data Center Manager | Sustainability: https://www.intel.com/content/www/us/en/software/sustainability.html?wapkw=sustainability

Follow Jake on Twitter at: twitter.com/jakesmithintel

Next Generation Data Center: The Future of Sustainability and the Cloud - CitC Episode 293

The SafeHouse Initiative was organized to provide education and awareness to businesses who are inundated with technical concepts, jargon and terms around the concepts of Business Continuity and Cyber Security. We understand that customers want to focus on their business rather than becoming Cyber Security or Disaster Recovery/Business Continuity technology experts.

Follow Jake on Twitter at: https://twitter.com/jakesmithintel

Safehouse Initiative Podcast - CitC Episode 294

Safehouse Initiative Podcast - CitC 294 Part 2

Safehouse Initiative Podcast - CitC 294 Part 3

Nelson Nahum, CEO and Co-founder of Zadara Storage, joins us to discuss Zadara's Intel-based Storage-as-a-Service (STaaS) products. Zadara's on-premises, hybrid, and cloud STaaS solutions eschew up-front capital expenditures in favor of more manageable, ongoing operating expenditures. Through this STaaS model, Zadara offers enterprise-grade performance, security, and control, while enabling customers to pay only for the storage they need. Nahum notes that Zadara allows the on premise storage to be easily moved and replicated to public cloud providers like AWS and Azure while maintaining the performance and security of the on premise environment. To learn more, please visit https://www.zadarastorage.com and follow Zadara Storage on Twitter at https://twitter.com/ZadaraStorage.

Zadara Offers Enterprise Storage-as-a-Service Using Intel Architecture - CitC Episode 68

Lee Howard, CTO of Global Industry Solutions and Alliances with NetApp, joins Conversations in the Cloud this week to discuss a shift in data management and the criticality of businesses leveraging their available data for successful decision making regardless of the platform where it resides.  NetApp’s vision is that the benefits customers have come to expect in a traditional on-prem architecture are protected as they implement a Cloud environment.  Howard discusses the goals that drive their data management strategy – create an agile environment that takes into consideration issues of data sovereignty and data resilience, while guaranteeing business continuity.  Visit cloud.netapp.com for use cases and resources, netapp.io for deep dive technical materials, and follow Howard on Twitter @KeepHoweird.

Understand the Power of Data Through Digital Transformation with NetApp -  Intel CitC Episode 135

In this episode of Conversations in the Cloud, Jacob Smith, SVP of engagement and Co-founder of Packet, joins us to discuss the promise of the cloud delivered on bare metal. Jacob and his brother founded Packet in 2014, with the intention of automating fundamental infrastructure. Jacob talks about Packet’s new collaboration with Intel to accelerate adoption of Intel® Optane™ SSDs as a data center fast storage and memory solution. Packet will provide on-demand access to the Intel® Xeon® Scalable platform with large format Intel® Optane™ technology at no cost. This will provide an easy way for open source and commercial software projects to benchmark and optimize this cutting edge technology, and unleash new innovations. To learn more about Packet and how you can accelerate with Optane, visit https://www.acceleratewithoptane.com/.

Hardware as the Next Innovation Layer: From Optane to FPGA's and Beyond- Intel CitC Episode 141

Intel® Conversations in the Cloud is a podcast with IT leaders who are driving the future of a software-defined infrastructure based data center. Featuring members of the Intel® Builders programs, Intel experts, and industry analysts, this recurring podcast series provides information on delivering, deploying, and managing cloud computing, technology, and services in the data center and enterprise.

Intel CitC

Popular tracks