NimbleEdge

?? Watch NimbleEdge’s session on Real-time Personalization in Mobile Gaming at Game Developers Conference 2024! We are excited to share NimbleEdge CEO, Varun Khare’s session from GDC 2024. In the session, Varun talks through the massive potential of real-time personalization in mobile gaming, and the enormous role on-device ML can play in this space?? Here are the key highlights from the talk: ?? Modern Mobile Gaming Landscape: The mobile gaming industry is evolving rapidly to become more data driven, with unprecedented user volumes. However, competition is increasing correspondingly ?? Importance of Machine Learning (ML) in Mobile Gaming: ML can be a pivotal differentiator in mobile gaming, with potential to significantly enhance user experience ??? Challenges in ML on Cloud: Most ML today is on cloud, which is slow, expensive, privacy-invasive, and often relies on stale customer data ?? On-Device ML and it’s use-cases: Enter on-device ML! By leveraging user devices for compute, real-time ML can be cost-efficient and privacy-preserving. This unlocks use-cases like real-time recommendations, offers, content moderation and more How NimbleEdge Helps: Despite its many benefits, optimizing real-time ML model execution across diverse devices is a massive challenge to take on in-house. NimbleEdge platform helps enterprises easily deploy and maintain ML models - users only have to upload the model, while the platform handles execution and orchestration!? Curious to learn more about how on-device ML helps with real-time personalization? Write to us at [email protected] #MobileGaming #MachineLearning #GDC2024 #AI? https://lnkd.in/gPZJhija?

Check out this recent blog by Google about on-device AI, which provides a clear overview of: ? What on-device AI really means ?? How today’s smartphone hardware powers on-device AI ?? The cost and latency benefits of on-device AI OS providers for smartphones and laptops have shipped on-device LLMs to deliver several valuable features. However, mobile apps still struggle to leverage the benefits of on-device AI due to the complexities of on-device model deployment and execution. At NimbleEdge, we’re solving this challenge with an on-device AI platform that enables edge modeling, deployment, event ingestion, model execution, and monitoring. Learn more at nimbleedge.com or reach out to us at [email protected]. ?? https://lnkd.in/efvEpBbJ

?????????? ?????? ???????????????????? ?????????????????????? ???????? ????????????, Arpit Saxena'?? ???????????? ???????? ???? ???????????????? ???????????? ???????????? ??? In the blog, Arpit breaks down hardware memory models and the complexities of relaxed concurrency, focusing mostly on ARM and IBM POWER architectures, while also motivating the C++ memory model ?? Ideal for developers looking to deepen their understanding of low-level memory synchronization, this blog offers valuable insights into ensuring correctness while squeezing out performance! https://lnkd.in/gS9_TqX7

We are stoked to welcome Neeraj Poddar to the NimbleEdge team as our new VP of Engineering! Neeraj brings a remarkable background in building infrastructure products for massive scale to NimbleEdge. He has previously co-founded Aspen Mesh and led the engineering team at solo.io, where he also spearheaded Istio, one of the largest and foundational open-source projects in the cloud-native ecosystem ?? Here, we share insights from a brief conversation with him, outlining the vast experience he brings to the organization and why NimbleEdge's vision resonates strongly with him ??? https://lnkd.in/gDB7A7ZH

?????????????? ???????????????? ?????? ???? ???????????? ?????????????????? ?????? ?????????????????? ???? ???????????? ?????? ??????????????! ?? French AI startup, Mistral AI, which builds foundational AI models, just launched its first models designed to be run on edge devices - Ministral 3B and 8B! ?? This launch heats up the competition in the sub-10B parameter language model category, with Mistral claiming its models perform better than similarly sized models by peers (e.g. Google's Gemma 2 2B, Meta's Llama 3.2 3B) across benchmarks?? While ML engineers now have a lot of choices in terms of edge-compatible models to build with, setting up and maintaining edge AI pipelines remain highly challenging, not least due to device diversity and performance issues on resource-constrained mobile devices. That is where NimbleEdge steps in. NimbleEdge platform simplifies the entire on-device AI lifecycle for mobile apps' ML teams, enabling effortless experimentation, deployment, execution, control and monitoring. Interested in learning more? Visit nimbleedge.com or reach out to [email protected] https://lnkd.in/gSCsGJFP

Watch this succinct summary of the on-device AI features already on offer in current flagship smartphones by BBC News! ?? With MediaTek announcing their new smartphone system-on-chip with dedicated NPU yesterday, it is clear that the possibilities with on-device AI will continue to expand rapidly? ?Mobile device manufacturers and operating systems are already capitalizing on this but performant on-device AI remains out of reach for mobile apps due to the complexity involved in edge AI experimentation, deployment, control and monitoring NimbleEdge is helping mobile apps tackle this problem with turnkey infrastructure for the complete on-device AI lifecycle! If this sounds interesting, visit nimbleedge.com to learn more or reach out to us at [email protected] https://lnkd.in/giw4hNYF

?????? ????-???????????? ???? ???????????????? ???????????????? ???? ?????????????? ???? ???????? ???????????? ? In a clear sign of the emerging importance of AI features in PCs, most coverage around the recent Windows update announcement has centred around new on-device AI features! ?? With this new update, Copilot+ PC users will be able to leverage Windows' on-device AI to: ?? Enhance the quality of old low-resolution photographs ??? Easily delete or generatively add elements to photographs ?? Access AI based features (e.g. object erasure, visual search) directly from the right-click menu on a snipped image In line with emerging on-device AI capabilities at an OS level, NimbleEdge helps mobile apps easily leverage on-device AI to enable truly real-time personalized experiences, using both traditional and generative AI. To learn more, visit nimbleedge.com or reach out to us at [email protected] https://lnkd.in/g39CPpVj

???????????? ?????????? ???????? ???? ????????????, ???????? ??????, ?????????????? ???????? ?????? ???????????? ???? ?????? ???? ???????????? ?? More and more AI use-cases are coming to the smartphone, such as text summarization, writing support, and translation ?? Hence, improvement in smartphone hardware for better on-device AI execution holds transformative potential. Just last week, we wrote about how NPUs are becoming increasingly common in smartphones to support on-device AI! ?? In this context, Jony Ive's reported new AI focused device is especially exciting news with increased competition in the AI smartphone space also likely to push smartphone incumbents to innovate ?? While on-device AI is now commonly employed by smartphone manufacturers due to latency, privacy and cost benefits, mobile apps still rely heavily on cloud for AI-ML use-cases. NimbleEdge helps users easily deploy, run and maintain on-device AI/ML, including session-aware and Gen AI use-cases! To learn more, visit nimbleedge.com or write to us at [email protected] https://lnkd.in/eezmpEsv

???????? ?????? ???????? ?????? ???????? ???? ?????????? ???????????? ???? ????-???????????? ????? ?? Neural Processing Units or NPUs have become increasingly important for electronic hardware manufacturers with the explosive rise in the popularity of on-device AI. But what exactly are NPUs? ? Put simply, NPUs are extremely fast processors, capable of performing trillions of operations per second, with provisions for parallel processing of vast amounts of data. Consequently, they yield significantly better performance in on-device AI tasks, at lower energy utilization than CPUs or GPUs! ? Historically, one key concern around deploying larger AI-ML models on-device had been resource constraints and performance issues, which could lead to sub-par customer experience. However, NPUs have alleviated this concern to some degree, which has been key for the advent of on-device Gen AI ? NPUs are already part of most recent flagship Android and iOS smartphones (e.g. Galaxy S21 and later, iPhone 12 and later), and will soon be present in the majority of smartphones, which is bound to be a massive boost to the continued growth of on-device AI adoption! https://lnkd.in/gK4DrPQx

Interested in learning about productionizing on-device AI models? ?? Watch this comprehensive overview by Instagram Product Manager Avni Kakkar. Going through the end-to-end on-device AI lifecycle, the video discusses: ?? ?????????????????? ???????????? ????-???????????? ????: Improved device hardware, better and smaller models, and customer demand for privacy-preserving AI have all contributed to the growth of on-device AI ? ?????? ???????? ?????? ????????: While on-device AI preserves user privacy and enables significantly lower costs and latency, there are limits to the size of models that can be run on-device given hardware resource constraints ?? ???? ??????????????????: The video goes through the various steps involved in achieving "ML readiness" for on-device AI pipelines, including model optimization, pre and post-processing, and evaluation ?? ???????????????????????? ????-???????????? ???? ??????-???????? ???? ??????????????????: Instagram uses on-device AI for it's "Cutouts" feature, which enables users to cut out objects from any photo in their camera roll. Given the sensitive nature of the input data, on-device AI is especially key to ensuring privacy! ? While the on-device AI ecosystem has grown significantly in recent times, cutting-edge high-value use cases such as Gen AI and session-aware AI remain difficult to implement on-device NimbleEdge platform helps mobile apps seamlessly tackle key challenges associated with session-aware AI and Gen AI with ease - be it resource consumption, device diversity or real-time data processing Curious to explore more? Visit nimbleedge.com or reach out to us at [email protected] ? https://lnkd.in/g9ERrSuB