Meet Team Qualcomm at AWS re:Invent
Las Vegas
November 27 – December 1, 2023
Booth #261
AWS re:Invent offers cloud enthusiasts, experts, and practitioners an experience to see innovations in real-time from cloud data center, robotics, gen AI, edge computing, IoT, automotive and much more.
Join for 5 days of opportunities, meet Qualcomm team and experience how to accelerate your AI workloads — from the cloud to the edge — with the power of Qualcomm platforms-optimized software, at-scale and efficiently.
Speakership highlights: Sriram Palghat, Director, Product Management, Automotive will highlight our collaboration with BMW to build a highly automated driving platform that leverages AWS cloud infrastructure.
SPEAKERSHIP:
How BMW & Qualcomm Technologies built an automated driving software platform on AWS
Developing highly automated driving features is a data-driven, compute -intensive process that includes a complex tool chain. From data ingest to algorithm training and large-scale simulation, learn how BMW utilizes AWS Cloud infrastructure for data management and tool chain integration to accelerate automated driving feature development for the new class of vehicles starting production in 2025.
SPEAKERS: Sriram Palghat (Qualcomm), Maria Guerra (AWS), Jörg Krebs (BMW)
November 30, 12:30pm
Session #: AUT202, Caesars Forum, Level 1, Summit 214
DEMO 1:
Qualcomm® Robotics RB5 AMR reference design with AI demo
This station demonstrates the edge processing capabilities of Qualcomm Robotics RB5 Development Kit paired with an open-source Large Language Model (LLM) hosted on the AWS cloud. The demo highlights how Qualcomm Technologies' suite of devices brings the best of the cloud and edge compute to robotics.
Based on user command in informal English text, the cloud based LLM determines the item best suited for the requestor. Once the LLM determines the suitable item, the robot moves and identifies where that item is and picks it up – all using on device Object detection, spatial estimation and robot arm control.
Using the power of AWS to perform natural language processing tasks while doing camera-based object detection and robotic arm control on the edge is just one example of how one can leverage the power of the cloud but perform privacy-sensitive or time-critical tasks securely on the edge.
Qualcomm® RBx devices with multi-modal connectivity, on device optimized compute and power efficiency are a great solution for state-of-the-art robotics!
DEMO 2:
Qualcomm Cloud AI 100 demos
With Qualcomm’s Cloud AI 100 inference accelerators now available in AWS EC2, we are showcasing a variety of demos powered by Qualcomm® Cloud AI. Qualcomm Cloud AI SDKs enable developers to optimize trained deep learning models for high-performance inference on Qualcomm Cloud AI accelerators.
Gen AI Demos:
Question and Answer - Powered by Llamav2 7B parameter LLM. Retrieval augmented generation optimizes the output of the LLM with enterprise-data to provide context-relevant results.
Text to Image – Generate images in less than 2 seconds using Stable Diffusion model that generates images from text prompts.
Transcription – Transcribe multiple audio streams simultaneously using Whisper Lite model that transcribes audio files into text.
Translation - Translate multiple languages simultaneously using "Opus" transformer-based models that can handle multiple language pairs.
Xgen - Chat demo with the XGen-7B model showcasing long context lengths and multiple tasks (including Q&A, translation, coding).
StarCoder - Code generation powered by a StarCoder 7B parameter LLM with support for 80+ languages.
Enterprise Security:
Kognition Demo - Kognition AI enterprise software running on Qualcomm Cloud AI 100 accelerator uses AI to interface with new and existing security cameras, access control systems, and IoT sensors to deliver an analytics platform with real-time alerting for smart buildings.
Try our sample apps on GitHub:
Learn how to run the Stable Diffusion text-to-image generative AI model
Convert and run language models
Convert and run computer vision models
DEMO 3:
BMW Automotive Demo
Accelerate autonomous driving with AWS and Qualcomm Cloud AI 100. In this demo, we are showcasing how AWS and Qualcomm Technologies are working to accelerate autonomous driving and software-defined vehicle transformation. ADAS developers can now provision EC2 instance with Qualcomm Cloud AI 100 to test advanced AD/ADAS capabilities with bit level accuracy and execute reprocessing at scale.
The Qualcomm Data and Simulation Factory Ecosystem is readily available to be deployed in AWS Ecosystem using AWS open-source frameworks like ADF and ADDF.
It is capable of processing 100s of PBs of data to help OEMs develop features, analyze performance, iteratively train ML models, and visualize data. With a commercial-scale data platform and extensive toolchains like the Snapdragon Ride™ Autonomous Driving Stack, we are working with developers to accelerate the overall time to market.
Attend our speakership “How BMW & Qualcomm built an automated driving software platform on AWS”
Join us on November 30 at 12:30pm
Session #: AUT202, Caesars Forum, Level 1, Summit 214
Snapdragon and Qualcomm branded products are products of Qualcomm® Technologies, Inc. and/or its subsidiaries.