Sessions for Developers

OEMs including CARIAD, Mercedes-Benz, Pony.ai, and TuSimple, plus Tier 1 suppliers Valeo and  ZF showcase the latest innovations and advancements in autonomous vehicles on a variety of topics including software, architecture, AI compute, sensing, and more.

ETH Zürich, Ottonomy, Ouster, and others reveal how they are building intelligent, safer transportation from the ground up, utilizing new advances in perception, high-performance sensing, and more.

BMW, Cruise, Microsoft, Stellantis, Zoox, and more discuss how autonomous vehicles are born in the data center, where high-performance compute solutions develop, train, test, and validate self-driving software.

BMW Group, DeepScenario, dSPACE, Luminar, Renault, and others discuss how high-fidelity, cloud-based simulation is accelerating autonomous vehicle deployment in the real world.

ESI Group, Hexagon, Maya HTT, Volvo Trucks Group Technology, and others showcase innovative ways to use GPUs for high-performance compute and graphics across transportation, manufacturing, and automotive.

> Chintan Shah | Senior Product Manager , NVIDIA

> Akhil Docca | Senior Product Marketing Manager, NVIDIA  

Learn how to create and deploy custom, production-ready vision AI and conversational AI models without any expertise in AI.

> Alvin Clark | Product Marketing Manager, NVIDIA

> Carlos Garcia | Product Manager, NVIDIA

Check out NVIDIA DeepStream SDK's newest features and how they can help you to get the most out of your vision AI apps.

> Stephen Chen | Applied Scientist, Microsoft

> William Raveane | Senior AI Engineer,  NVIDIA

Learn how Bing worked with NVIDIA to accelerate inferencing parts of their Visual Search AI features and meet challenging constraints in terms of latency and accuracy.

> Brian Duffy | Director of Marketing,KLA

> Jerry Chen | Director of Global Business Development, Manufacturing & Industrials, NVIDIA

Learn the best practices used to address challenges of defect detection in semiconductor manufacturing while maintaining a reasonable computational infrastructure cost.

> Guy Elad | Deputy CIO, Tel Aviv Municipality

> Clayton Garner | CIO, City of San Jose

> Michael Sherwood | CIO, City of Las Vegas

Hear from industry leaders how they’re solving modern challenges with AI in computer vision and providing a positive impact on smart spaces.

> Bartley Richardson | Senior Data Scientist and AI Infrastructure Manager, NVIDIA

> Hema Vivekanandan | Security Analyst, NVIDIA

Learn how you can take advantage of the latest AI and data science tools and frameworks, pre-trained models, and infrastructure optimization software to tackle cybersecurity challenges in new ways.

> Nick Tobash | Director, Incident Response, Cyber Hunt and Investigations, Best Buy

Join Best Buy's security experts as they talk about how they're integrating NVIDIA Morpheus into their data science pipeline to increase security through network mapping and phishing detection.

> Dorit Dor | VP Products, Check Point

Join this session to learn how to use artificial intelligence and automation to prevent the most sophisticated cybersecurity threats.

> Sunil Vasisht | Staff Data Scientist, Mandiant

Join this session to learn how traditional disassembly can be framed as a neural machine translation task, leading to a more flexible approach while significantly reducing the overhead required to perform such person-intensive tasks.

> Úlfar Erlingsson | Chief Architect, Lacework

Learn how Lacework and NVIDIA are partnering to bring data-driven security to large cloud infrastructures, leveraging a natural language approach to cybersecurity anomaly detection.  

> William Benton | Principal Product Architect, NVIDIA

> Sophie Watson | Technical Product Marketing Manager 

Savvy practitioners love to dive deep into machine-learning techniques, but they know that even the most elegant technique isn't the whole story. To get business value from AI, you'll need to build a machine-learning system, which means coordinating cross-functional teams, polyglot services, and diverse data sources. Learn how NVIDIA AI Enterprise, NVIDIA EGX, and the RAPIDS Accelerator for Apache Spark provide benefits across an entire machine-learning system: accelerating discovery, model training, and model serving, feeding a feature extraction pipeline tens of times faster by running Apache Spark queries on Ampere GPUs and making all of this power accessible.

> Khalifeh AlJadda | Senior Director of Data Science, The Home Depot

Recommender systems play a crucial role in any ecommerce platform. Therefore, it's very important to build a personalized, scalable, and reliable recommender system that proactively helps shoppers discover products that engage them while on the website, and then re-engage them with products and content that align with their interest after they leave the website via email, social media, or other marketing channels.

I'll cover our transformation at The Home Depot to an AI-based recommender system leveraging the power of deep learning and GPU. We'll discuss how we train and deploy real-time deep-learning models to enable personalized recommendations at scale.

> Sirisha Rella | Product Marketing Manager, NVIDIA

Conversational AI technologies are becoming ubiquitous, with countless products taking advantage of automatic speech recognition, natural language understanding, and speech synthesis coming to market. Thanks to new tools and technologies, developing conversational AI applications is easier than ever, enabling a much broader range of applications, such as virtual assistants, real-time transcription, and many more. The talk will be composed of two parts. We'll start with an overview of the conversational AI landscape and discuss how any organization can start developing conversational AI applications today. Then we'll demonstrate how to build and deploy end-to-end conversational AI pipelines using NVIDIA Riva.

> Miguel Angel Martinez | Senior Deep Learning Solution Architect, NVIDIA

> Meriem Bendris | Senior Deep Learning Solution Architect, NVIDIA 

Recent advances in natural language processing demonstrate the capability of large-scale language models (e.g. GPT-3) to solve a variety of NLP problems with zero shots shifting from supervised fine-tuning to prompt engineering/tuning. However, building large language models raises challenges on data preparation, training, and deployment. In addition, while the process is well-established for a few dominant languages such as English, its execution on localized languages remains limited. We'll give an overview of the end-to-end process for building large-scale language models, discuss the challenges of scaling, and describe some existing solutions for efficient data preparation, distributed training, model optimization, and distributed deployment.

> Shankar Chandrasekaran | Product Marketing Manager, NVIDIA

> Mahan Salehi | Product Manager, NVIDIA

We'll go over NVIDIA Triton Inference Server software and what's new. Triton is an open-source inference-serving software for fast and scalable AI in applications. Learn how Triton helps deploy models from all popular frameworks—including TensorFlow, PyTorch, ONNX, TensorRT, RAPIDS FIL (for XGBoost, Scikit-learn Random Forest, LightGBM), OpenVINO, Python, and even custom C++ backends. Also learn about the features that help optimize inference for multiple query types—real-time, batch, streaming, and model ensembles.

> Alexey Panteleev | Distinguished DevTech Engineer, NVIDIA

NVIDIA provides three SDKs for real-time lighting and denoising. RTXDI implements a set of resampling algorithms for computing direct lighting and ray-traced shadows from many lights—up to millions—without baking. RTXGI approximates indirect lighting using grids of dynamically updated, ray-traced light probes. Finally, NRD provides a collection of real-time denoising algorithms in an easy-to-integrate package.

> Lina Halper | Principal Animation Engineer, NVIDIA

Animation retargeting has been a challenge for many years while the demand of diverse characters and use cases are increasing tremendously outside of the traditional gaming industry, as in synthetic data generation or digital twin, for example. Downloading animations online is easy to do, and there is much research on generating those animations automatically, but applying them to a custom character can be challenging and requires a huge technical understanding of rig, skeleton, and animation. Also, there are very limited open options in the industry, and the quality ranges from low to high, with many knobs and tweaks required from users. We'll explain how we tackled this in the Omniverse while aiming for an easy-to-use feature.

> Andrew Edelsten | NVIDIA

NVIDIA Deep Learning Super Sampling (DLSS) uses the power of deep learning and AI to increase game frame rates through the clever use of spatio-temporal upsample and the NVIDIA Tensor Cores. During the session, the speaker will walk through the underlying technology that makes DLSS possible, and also explain how to integrate DLSS into a new game engine. Learn about the engine requirements for DLSS and things to look out for during an integration. Finally, the session covers general DLSS debugging tools and how to fix common issues that may be negatively affecting the image quality of the final frame.

> Markus Gross | Vice President of Research, Walt Disney Studios

> Simon Yuen | Director of Graphics and AI, NVIDIA

> Matt Workman | Developer, Cine Tracer

> Sarah Bacha | Head of Research and Innovation, Cedrus AI

 > Vladimir Mastilovic | VP of Digital Humans Technology, Epic Games

Join technical and business luminaries for this panel discussion exploring how digital humans will have a profound impact on industries outside of Media and Entertainment. We’ll explore what leading technology companies are contributing to the space today, and what they see for the future.

> Peter Morley | Senior Developer Technology Engineer, NVIDIA

Ray tracing is primed to replace rasterization due to advantages in illuminating environments providing more realistic and immersive experiences in games. Open-world, ray-traced games typically employ hybrid graphics renderers that take the best pieces from ray tracing and rasterization. This talk will primarily focus on the performance and scalability of ray tracing when converting a raster-based renderer into a hybrid (rasterization and ray tracing) renderer.  Specifically, we mention acceleration structure building approaches, ray tracing regimes that scale efficiently, and ray tracing workload scheduling to get the best bang for your buck on NVIDIA RTX hardware.

> Damien Fagnou | Senior Director of Software, NVIDIA

Learn to easily extend the Omniverse platform by creating extensions—the building blocks of NVIDIA Omniverse Kit-based apps such as Omniverse Create or Machinima. Be sure to join “Developer Breakout: Building Your Own Extensions and Apps on Omniverse Kit” as well to talk with Omniverse experts!

> Lou Rohan | Omniverse Connect Engineer, NVIDIA

Learn how to connect with the Omniverse platform and be able to send data to it, establish a live sync session, and get a USD 101 overview to get you started.

> Gavriel State | Senior Director of Simulation and AI, NVIDIA

Join a deep dive into building microservices with Omniverse and how to use and run them. Be sure to join the “Developer Breakout: Build Your Own Microservices on Omniverse” as well to talk with Omniverse experts!

> Michal Hapala | Senior Engineer, NVIDIA 

> Kier Storey | Distinguished Engineer, NVIDIA 

> Michelle Lu | Director of Simulation Technology, NVIDIA 

> Adam Moravanszky | Director for Simulation Technology, NVIDIA 

> Ales Borovicka | Senior Manager, Omniverse Physics, NVIDIA

We'll take a deeper dive into building microservices with Omniverse and how to use and run them.

> Miles Macklin | Principal Engineer, NVIDIA

Join this presentation on Warp, a runtime framework that allows running Python kernel code in CUDA, and its integration into Omniverse.

> Andrea Thomaz | Co-Founder and CEO, Diligent Robotics

Dr. Andrea Thomaz will discuss the unique challenges in achieving robot autonomy in a dynamic human environment like busy hospitals–from having to maneuver around objects in busy halls to having to navigate to a requested location, all while integrating multiple camera streams that feed into embedded GPUs. View the full Robotics playlist.

> Dieter Fox | Sr. Director of Robotics Research, NVIDIA

This in-depth session covers recent work on learning to manipulate unknown objects directly from visual (depth) data and combining pre-trained language and vision models in order to teach a robot to perform manipulation tasks. View the full Robotics playlist.

> Robert Banfield | Machine Learning Engineer, Trimble

> Nyla Worker | Solutions Architect, NVIDIA

This session explores the generation of deep learning models to achieve state-of-the-art predictive accuracy by combining publicly available and synthetically generated and labeled point clouds. View the full Robotics playlist.

Jetson Developer Day is a special event open to developers of all levels. Featuring deep-dive sessions with Jetson platform experts, this collection of talks showcases the latest innovations in autonomous machines and edge AI.

> Raffaello Bonghi | Developer Relations, NVIDIA

Get a technical overview of how NVIDIA hardware-accelerated robotics software packages—Isaac ROS GEMs—are used to help a robot's perception, localization, and mapping. View the full Robotics playlist.