Mushan Yang

How you can make your LLM applications better!

5

🎉 Big congratulations to our incredible team (Haoyu Chen, Nick Greenquist,Yuxing Wang, Sridevi Maharaj, Ruiyang Wu and many more) for delivering several quarters-long impactful features on the Play Store! 🚀 We’re thrilled to have launched key updates like Search Home, Sports Moment, and WearOS recommendations, all designed to enhance personalization and help billions of users worldwide search and discover content they love. 🌍✨ Check out the Play Store Q3 updates: https://lnkd.in/dn5wx8Mi #GooglePlay #Personalization #Recommendation

74

1 Comment

Come join us tomorrow in SF to learn more about the latest updates to run AI workloads on Google Cloud!

11

1 Comment

I recently had the opportunity to attend 3 major AI events in the last 30 days, where I gained valuable insights into the future of artificial intelligence. Here are some key takeaways: Amazon AI Day: - Amazon CTO (Werner Vogels) highlighted the potential of generative AI and large language models in driving innovation and societal impact, particularly in healthcare and sustainability. - Explored how generative AI is revolutionizing development processes and transforming businesses across industries. MagicBall - AI Week: - Showcased advancements in AI infrastructure by over 24 startups from Bengaluru, emphasizing the city's significance in AI development. - Engaged in discussions on responsible AI development and ethical considerations for the future. Microsoft AI Day: - Microsoft's Innovation Hub India CTO (Sandeep Alur) discussed AI's critical role in digital transformation and the company's commitment to ethical AI development. - Showcased the revolutionary potential of Generative AI and its integration into Microsoft Azure. These events, combined with insights from the ISB CTO program, emphasize that AI is here to create new opportunities and possibilities, not to take away jobs. Embracing a growth mindset is crucial to thrive in the evolving AI landscape. The AI revolution has just begun, and it's essential to adapt and innovate for a sustainable future. #GenAI #AI #BrightFuture

29

6 Comments

For the first time I am happy about certificate which I received today that remind me the new learning, foundation for next year, and the hours I spent over the weekend for the 10 weeks. Little bit about course: NLP with deep learning The best part is the hands on exercises where we learn the foundations of deep learning, neural networks and the NLP. But it will be useful as a starting point and set the direction in that path.

50

10 Comments

Today I'm sharing this 🔥 video from our recent Engineering onsite at Demandbase HQ in SF! https://lnkd.in/eFAFR6QG We were so fortunate to host Zhengxuan Wu and Aryaman Arora, two upcoming AI researchers from Stanford University Department of Computer Science. Their approach to fine tuning, called ReFT, targets the representations emitted by neural net layers rather than the weights of the layers themselves. This makes it very flexible when using the same base model for swapping multiple fine tunes. The memory overhead is also very low for their intervention weights. Another highlight is how powerful it can be. One POC they showed forced the LLM to speak in only emojis exhibiting ReFT's power to control output. Control is crucial when, for example, you want to prevent your bot from offering customers unauthorized refunds!! If you're already familiar with LoRA, it should be pretty simple to use with only a few extra hyperparameters -- details in the video. Here's a link to the repo. Give them a star! https://lnkd.in/eJScNbzK Thanks again for letting us host your first external talk Aryaman and Zen. It's very exciting tech!!

17

4 Comments

Getting started with AI can seem daunting, especially when contemplating its safety against data privacy. But past that fear lies a fix for itself: AI tools can be harnessed to overcome cybersecurity hurdles. Join us for an insightful discussion at the intersection of security, building resilience, and emerging technologies like AI, IoT, and others such as operational technology, and products intersecting with information technology and engineering. Get inspired, learn something new, and connect with peers, partners, and Googlers!

5

Watching Professor Girotra talk about recent developments in Chinese companies using GenAI and putting that in perspective with what we learn from Western companies, it’s exciting to see the technology’s potential being explored from different angles. Take a look at Kling, the GenAI tool that Kuaishou (a company comparable to TikTok) built to allow its users to generate videos up to two minutes long with a resolution of 1080p / 30 fps. This video is from a prompt: ‘A giant panda is playing guitar by the lake.’ Girotra's video - https://lnkd.in/gfanV4Xk. He's also the author of the Generative AI for Productivity certificate program published by eCornell. #IWork4Dell #GenAI

11

I'm excited to host the AI4D: ByteDance Open Source Meetup on Thursday, October 3rd, from 5:30 PM to 8:30 PM at 1199 Coleman Ave, San Jose, CA 95110. Join us for a gathering of AI professionals to discuss challenges and solutions in AI implementation. Topic: AI/ML Inference Our keynote speaker is Dr. Alex Wei Chen, Founder and CEO of NEXA AI and creator of Octopus v2, an AI model that's 4x faster than GPT-4 and 70x more energy-efficient. He'll introduce Nexa SDK, an on-device AI toolkit supporting various models and GPU acceleration. Agenda: - 5:30 PM - 6:00 PM: Networking & Registration - 6:00 PM - 6:30 PM: Welcome and AMA with Dr. Alex Wei Chen & Tina Tsou - 6:30 PM - 7:30 PM: Nexa AI Presentation - 7:30 PM - 8:30 PM: Networking Dinner and refreshments will be provided. Space is limited, so RSVP is required. Please remember to bring your photo ID. If you're interested in presenting your company's product, contact Gary at garyzhen@coze.com. Follow ByteDance Open Source on Twitter: [@ByteDanceOSS](https://lnkd.in/g5PDdYyY) Looking forward to insightful discussions and networking with leading minds in the AI industry! #AI #MachineLearning #OpenSource #Networking #ByteDance ByteDanceOSS

33

In the AI ecosystem, LLMs act like CPUs, crucial yet dependent on complementary tools such as Chatbot and AnythingLLM, which function similarly to I/O systems. Alongside, tools like MemGPT and CrewAI serve roles akin to RAM and network interfaces. The challenge arises as these tools often evolve independently, aspiring to be standalone solutions. This siloed development leads to a lack of standardization, complicating their integration and limiting their utility in professional settings. Without unified protocols, integrating these diverse AI tools with LLMs demands significant customization, placing an undue burden on developers and hindering broader innovation. Just as the adoption of standardized interfaces like USB revolutionized computer usability, AI tools would benefit from similar standardization to enhance functionality and ease of use. Promoting common standards for AI tools is crucial for maximizing the capabilities of LLMs and ensuring their practical application across various domains. Without such efforts, the AI field may continue to struggle with fragmentation, undermining the potential of these advanced technologies.

3

2 Comments

https://lnkd.in/gxJWb4hA Just testing! :) I "wrote" this short article with help of AI - in minutes, but it is about an important difference, that I would like to highlight, between both awesome APAC hubs: HK and SG. For example it would be great if HK introduces LTVP visa for non-immediate family members and it would be also great if SG allows spouses to take up jobs again. The issue is complex, but worth a read!

6

Check out the latest episode of our Tech Teams Today podcast with Yongsheng Wu, VP of Engineering at Granica - now available on all major podcast platforms. #AI #EngineeringLeadership #DataScience #MachineLearning #TechTeamsToday #Revelo

4

I am excited to share the latest blog, co-authored by myself, Junhwi Kim and Vyas Sekar, detailing how we managed to successfully localized and improved LLM by combining Fine-Tuning and RAG. https://lnkd.in/gKDkgGPq

70

3 Comments

Thats the exact explanation I gave to a group of investors last week. (… it’s just Precision/Recall, attributes vectorization as “embedding”, K-Nearest Neighbor, User Search Intent, DAG, ML algorithms, etc all over again …). LOL and they asked what’s new this time around. And I was like not much, ML is still the backbone, and AI is the inference and actions based on the model. It works better when there is a feedback loop in a closed system. But the current state of the art like ChatGPT is based on scraping the Internet (garbage in, garbage out). Kinda like early search engines. What is a better analogy is that NVIDIA/Snowflake etc are selling tools, aka pickaxes to prospectors. The gold is up to you to find. And the opportunity right now is to monetize industry specific (healthcare, travel, financial services) clean data to build models with (aka a treasure map for the prospectors). From an enterprise perspective, what’s out there today aka generative AI isn’t very useful. Too much hype. To build useful engines will require an effort similar to search revolution 2 decades ago. From

16

3 Comments

Every leader should watch this video. Thanks for sharing Nolan Church! In his book Loonshots, Safi Bahcall breaks down an innovation equation. One of the reasons I love this so much is that it recognises one of my favourite quotes from the late Charlie Munger about the power of incentives: “Show me the incentive and I will show you the outcome.” The innovation equation explains what creates innovation and it revolves basically around one idea. If employees gain the most value personally by delivering outputs which benefit the organisation, you’ll have high innovation. If they gain more value by doing anything else, you’ll have lower innovation as they will prioritise those things instead. Equation: Innovation rate = M = (E + S^2 + F) / G Equity Fraction (E) – how much equity does the team have in the success of an idea. This can be tangible (i.e. stock) or intangible (i.e peer recognition). More equity = more innovation. Fitness Ratio (F) – how well employees skills are paired with the work they do. Strong pairing means people are aligned to delivering value through their work (high innovation) rather than by lobbying for favour and playing politics (low innovation). Management Span (S) – The number of direct reports managers have (higher = more innovation, lower = less innovation). Salary Growth (G) – how your salary grows as your rank increases (higher = less innovation). If you look at this equation through the case study of Nvidia (undeniably one of the more innovative companies of our time) you see this playbook being run. - Low level of hierarchy, the Ceo has 60+ direct reports - Generous equity compensation with life changing outcomes for employees - Heavy focus on fitting skills to jobs Running these through the equation, you’d expect a highly innovative business! I love finding one element in a system which can drive an outsize gain in outputs. Especially if it’s easy to do. In this case, implementing a low hierarchy organisational structure is that one element. Taking out hierarchy and expanding the number of direct reports managers have, will dramatically increase the focus on outputs while reducing politics! This in turn generates significant value for the organisation because it focuses everyone on doing what is actually important. That this is literally the opposite of what most companies currently do, makes me like it even more. 🙂 Summary article on the innovation equation from the Harvard Business Review here: https://lnkd.in/eiNKH7fQ

19

1 Comment

The challenges of scaling and recovering AWS MSK clusters are now effectively mitigated, representing a significant advancement in cluster management.

10

Why did Kong Inc. choose Cloudsmith for artifact management? Here’s what Saju Pillai, Senior VP of Engineering at Kong, had to say: 🗨️ ‘Our engineers love Cloudsmith. The complaints about our distribution mechanisms have gone to zero. We no longer need to worry about software distribution challenges.’ For a global API leader like Kong, every line of code matters. Its software distribution powers some of the world’s most essential services, from banks, to airlines, to e-commerce giants. Every artifact matters when uptime, security, and customer experience are on the line. That’s where Cloudsmith comes in. With 600+ global points of presence, seamless CI/CD integration, and fully managed artifact delivery, we help teams like Kong eliminate artifact chaos and focus on building extraordinary software. It’s what drives us every day: supporting innovators who keep the world connected, secure, and moving forward. 📖 Curious how Cloudsmith transformed Kong’s software delivery? Read the full story here: https://lnkd.in/e-XT7eJi #ArtifactManagement #SupplyChainSecurity #DevSecOps #Cloudsmith

21

1 Comment

LLMs and LangChain: The Dynamic Duo 🦸‍♀️🤖 Just finished a fun project where we built a Q&A bot using these two powerhouses. It's like having a personal AI assistant that's always ready to answer your burning questions! 🔥 Here's what I learned: LLMs are wizards: They can conjure up answers from thin air (or, uh, the internet). 🧙‍♂️ LangChain is the glue: It holds everything together and makes it easy to use. 🔗 Endless possibilities: This tech can be used for anything from customer support to writing catchy social media posts. 📝 Want to see it live (obvio a test server link 😛 ): https://lnkd.in/dZNAVCSV Note: it's a render link, they will take down the instance because of no use, you can hit the link and wait for two minutes, they will spin up the instance and you can try it out. Want to see the code? Check it out here: https://lnkd.in/dzjh4r47 #AI #MachineLearning #NLP #LangChain #LLM #Python #Developer #FunProject"

22