Jeremy Manson

Wrapping up my first week as a Headstarter AI SWE Fellow, I had the opportunity to attend the Pathways to Tech event where industry leaders shared invaluable insights. Here are my key takeaways: 🔍 Ask quality questions: Great questions leave a lasting impression. Remember, people may not remember you, but they will remember your great questions. 🗣️ Effective communication: Clear communication is a powerful skill. It's essential to clearly convey your ideas to your team and potential clients, as emphasized by Dustin Beadle. 🎯 Identify pain points: User retention hinges on solving their specific pain points, a key insight highlighted by in the event. 🔄 Adaptability: Shariar Kabir's journey from medicine to computer science reveals the importance of adaptability and continuous learning, showcasing the value of keeping an open mind to new areas of curiosity. 🌟 Diverse skill set: Continuous learning and a diverse skill set were underscored by Miguel Acero as essential for professional growth. 💼 Soft skills matter: Communication and community involvement are crucial for building relationships, fostering teamwork, and creating a positive work environment alongside technical skills. 💪 Resilience: Nabeel Alamgir🦄's story of overcoming rejection emphasizes the importance of resilience and staying strong in the face of challenges. 📊 Customer feedback: Actively seek and incorporate customer feedback for improvement early and often. Huge thanks to Yasin Ehsan 🚀, Faizan Ahmed, and the entire Headstarter AI team for organizing such an enriching event. Gratitude to all the speakers for their valuable time and lessons. Looking forward to applying these learnings in my future projects and career! If you're interested in watching the event, here is the link: https://lnkd.in/ePkMvN4n #HeadstarterAI

3

For anyone interested in Confidential Computing on Arm platforms, you must watch the update that Dan Handley from Arm provided at Linaro Connect. He covered the current state, what's coming next in the Confidential Compute Architecture (CCA) specification, and what to expect. Watch the recording at https://lnkd.in/dRiCz9TH.

24

Get ready! Our crew is setting up for live coverage of Anyscale #RaySummit, kicking off today! 👉https://lnkd.in/gQ8vEghU Join analyst-led conversations to dive deep into Anyscale’s plans for #ScalableAI under its new leadership. Learn about how #Anyscale plans to address AI infrastructure challenges and how it's leveraging strategic partners, such as Nvidia, to further its growth.✨ “LLMs are at the heart of today’s generative AI transformation, and the developers creating and customizing these models require full-stack computing with efficient orchestration throughout the AI life cycle,” said Manuvir Das, VP of enterprise computing at NVIDIA. “The combination of Nvidia AI and Anyscale unites incredible performance with ease of use and the ability to scale rapidly with success." Don't miss our exclusive coverage! #EnterpriseTech #LiveNews #CTOanalysis #MarketInsights #AIinfra

5

Vara Bonthu and the team continue to produce incredible assets, driving forward the extraordinary partnership that empowers customers to scale Large Language Models (LLMs) more efficiently on Amazon EKS with NVIDIA's Inference Microservices (NIM) Today’s spotlight is on how this collaboration makes deploying Meta's Llama-3-8B model on Amazon EKS simpler and more effective. With dynamic scaling supported by Karpenter, the solution ensures optimal performance and resource efficiency. Key highlights include detailed guidance on deploying, managing, and scaling LLMs, plus insights into leveraging Prometheus and Grafana for complete observability. Whether your focus is low-latency or high-performance inference, this approach lays the foundation for scalable, powerful AI model serving. Check out the AWS blog to dive into the full details and see how Vara’s leadership is helping customers fully leverage the power of NVIDIA NIM on EKS. Abhishek Sawarkar

15

1 Comment

Yes, Linux 6.11 comes with its own Blue Screen of Death (Well sort of purple actually, but you can get in in black as well), but there are far more important features in this latest Linux kernel release.

11

2 Comments

🌟 Pathways to Tech Event Highlights 🌟 I attended the Pathways to Tech event, and it was amazing! Here are some key points: - Shariar Kabir: Talked about the importance of innovation and hard work. As CEO of Ruby (YC 23) and ex-Salesforce, his advice was great. - Miguel Acero: Shared his journey from Google to CTO at Ruby, focusing on lifelong learning. - Dustin Beadle: Gave insights on leadership from his time at AMD and Lattice. - Nabeel Alamgir: Showed how resilience led to raising $100M for LunchBox. My favorite speaker was Shariar Kabir for his forward-thinking approach. With Ruby's new partnership, I plan to use this Fellowship to improve my skills and increase my chances of joining Ruby. 🚀 Missed the event? Watch it here: [Pathways to Tech Event] (https://lnkd.in/dXVyVpt9) #Networking #PersonalBranding #TechCareers #PathwaysToTech #Ruby #FellowshipOpportunities

1

Everything else being equal, finding bugs statically is better than finding bugs at runtime. Everything else is rarely equal. In the comments to my last post, someone asked about the benefits / adoption of formal verification compared to testing. I have some-but-not-a-ton of experience with formal verification, but I have a lot of experience with static analysis tools. For those of you who don't know, formal verification involves making a mathematical specification for what your code is supposed to do, and then running a tool that can prove it meets that specification. Most people reading this post will now know why most developers don't do it. (It's worth a reminder - tests *are* a specification for your code. They're just one that's written in the language in which you wrote the test instead of being written in math.) It's a truism that it is better to find a bug at compile time than at test time, at test time than at code review time, and at code review time than at deployment time. Each additional step bakes in problematic decisions, giving them the chance to spread to other code, and, eventually, to users. Also, every second that passes between when you write the code and when you notice the problem is a second your brain might use to forget what the code does. I wrote a post a couple of months ago about research done at Google to identify what the ideal static analysis tool looks like. In short: you don't have to do any work to make it run, it never gives you false positives, it points out real problems, and the tool tells you exactly what you need to do to fix the problem. Tools like this are great, and are why Google leans heavily on govet, clang-tidy, and Error Prone (errorprone.info). The more typing that someone has to do to make a static analysis tool work, though, the more that they are likely to write a test, instead (possibly adding a dynamic analysis tool). Formal verification tools are at the far end of this - they tend to make you write specifications for your code, which they then prove the code obeys. Everyone who has worked on a codebase in the broader tech industry knows that most specifications last (at most) as long as it takes to write the code, at which point you realize you got them wrong and change them. This makes formal verification a lot of work for very little reward. The only people who write formal specifications for their code are the ones who are writing code that will kill you if they get it wrong - code to keep airplanes flying or to keep nuclear reactors from meltdown. Most Java development is way back at the other end, where we still beg and plead with people to add nullness annotations (it will help when the jspecify.dev project finishes). This is why sneaking static analysis features into programming languages is so important (I hear you, Kotlin fans). At some point, people's interest in typing annotations drops off, and for everything beyond that, we have tests.

36

10 Comments

In a wide-ranging conversation, Jensen Huang, NVIDIA's CEO of 31 years, shared profound insights into leadership, innovation, and the transformative power of AI. Huang emphasized persevering through "pain and suffering" to achieve greatness, as nothing truly great was ever easy. He underscored the importance of deeply understanding first principles and having unwavering conviction in one's core beliefs when reasoning through challenges. Huang discussed NVIDIA's decade-long journey with CUDA, a pioneering technology that initially crushed the company's margins and market cap before becoming vital for modern AI. This experience taught him the value of being "clever" to sustain challenging endeavors and embodied his belief that achieving the extraordinary requires more than just operational excellence - it demands "craft" transcending metrics. Huang views AI as sparking a new industrial revolution, producing valuable "tokens" of intelligence that will reshape industries and drive productivity. He stressed aggressively embracing AI, warning "if you're not engaging AI actively and aggressively, you're doing it wrong." However, he advocated combining general AI capabilities with domain-specific models tailored to each organization's proprietary needs. Central to Huang's philosophy is nurturing talent through "tenure" to embody the collective wisdom, experience, and "encoded magic" that allows small teams to achieve outsized impact. This manifests in NVIDIA's flat structure with over 60 direct reports, fostering open information sharing and collective problem-solving. Huang believes in "torturing" talented individuals into greatness rather than firing them, exemplifying his commitment to reasonable stubbornness over obstinance. Ultimately, Huang attributes NVIDIA's $2 trillion market cap to prioritizing "beauty," "elegance," and "simplicity" that, though hard to specify, resonates with customers. He advocates for young entrepreneurs embracing this mindset, leveraging AI's potential while upholding core principles of craft and philosophy encoded by "tenure" over times of immense challenge. #nvidia #huang #jensenhuang #cuda https://lnkd.in/gJgCxmvU

1

Two years at Intel Corporation today, just like that! Five reasons I believe in Intel: 1. Pat Gelsinger is turning the company around. He is bringing the rigor, discipline, and execution and setting us on a path to success. His drive, vision, and infectious enthusiasm are steering us towards a bright future. 🚀 2. The culture at Intel around collaboration, innovation, and constant learning is just amazing. There are trusted tools and methodologies used for decision making which brings much clarity. Our meetings transcend multiple BUs and we focus more on task conflicts instead of personality conflicts. Employees are encouraged to take classes on diversity and inclusion with guidance that can be applied to work directly. 🌟 3. I love what I do. As an athlete, I challenge myself in workouts every day. At work, I love pushing the boundaries of open source. The management enables and supports me to chop-wood-carry-water in multiple open source communities, it has been our legacy for over two decades. It allows you to be your authentic self and is such a fun experience. 🌐 4. I love my team - a big shoutout to Shirley Bailes Jessica Marz Ryan Ware Rachel Roumeliotis David Wiltz Michelle Myers - you rock. My management's support and encouragement only helped me to be a better person — shoutout to Melissa Evers Jeff McVeigh Greg Lavender for being the wind beneath my wings. 💖 5. Intel puts "silicon" in Silicon Valley. I want to be part of the turnaround of such an iconic company, and help make it! 💡 This article (https://lnkd.in/g8AiwUSH) from The New Stack talks about how an open ecosystem creates a leveled playground and why walled gardens restrict innovation. This is what I enjoy working on every day, every moment. Here's to two years of growth, innovation, and sheer awesomeness at Intel! 🎉 🎊 🙌

110

8 Comments

This overview of Generative AI and Retrieval-Augmented Generation by Gwen (Chen) Shapira of Nile Database is fantastic. Highly recommend it for other infrastructure nerds like me who want to know what's going on in the AI world. #generativeai #rag https://lnkd.in/gJum7_qa

10

1:40 - that’s the ratio I advise between engineering leaders and their followers. This video calls out one of the biggest reasons leaders have for keeping their follower count way too low - one-on-ones. An assumed piece of “common wisdom” that’s simply not helpful. Hierarchy is a problem definitely - but the overly small team also makes the leader too tactical. It rewards micromanagement with results. And micromanagement is death to a software team. If you have more than 40 engineers and your engineering leaders have fewer than 40 direct reports, it’s time to consolidate.

3

https://lnkd.in/gFWsZBEM Keeping pace with LLVM: compatibility strategies

9

2 Comments

How Ray offers scalable compute to MLOps practitioners 🧑‍💻 Join #theCUBE for this Anyscale Compute’s #RaySummit highlight, where we’re hearing from Christian Stano, engineering manager of MLOps at Attentive, on how Ray Distributed gives MLOps practitioners access to a scalable, streamlined, and reliable compute platform. “There's an S curve of growth for machine learning and AI platforms where you have a lot of strong early results at the bottom of the S curve. You start growing into the growth phase and then you hit the data and the compute ceiling. The beauty of Anyscale’s Ray is that it unifies, from a data ingest point to inference output point, that whole ecosystem and framework with plug and play modules all over the place,” Stano shares. “For me as a MLOps practitioner, this is an absolute dream come true. From the data side, we can now lift all the ceilings that we've had for our data. From the compute side, we now have access to scalable compute clusters. That's where Ray comes in to give us access to that scalable compute platform and makes it easy, streamlined, scalable, and reliable,” he adds. 📺 Get the full story: https://lnkd.in/gstHyKbg

9

It is a hot topic about extending the measurement chain of trustworthy from launch time, OS runtime to cloud native in confidential computing environment, so the attester or verifier can choice the particular raw TCG event log for the next level attestation which defined by IETF. We (Wenhui Zhang Haidong Xia jiewen yao) developed "confidential cloud native primitive" (https://lnkd.in/g6_kgqTX) to collect the trustworthy measurement in container, namespace, cluster level. Please check our demo video at: https://lnkd.in/gwAEz9XT

3

1 Comment

GPUs are expensive and in short supply. Learn how to automagically consume just-in-time rightsized NVIDIA GPUs for your Ray clusters from Anne Holler and Justin Willoughby Anyscale #mlops #gpu #ray #nodelesskubernetes

20

What a special moment! Been working on this product with other talented engineers for few years now, and I am thrilled to share with you “Orion” our first Gen AR product that demonstrates the potential of this next gen compute platform! While still in the early innings, it's already showcasing the immense potential of this technology and the real utility it can bring(checkout the link to the demoes below). Building products of this complexity takes immense level of innovation, partnership and architectural trade-offs across custom silicon, sensors, ID, thermals, PDE and EE miniaturization. The technical challenges are steep, but moments like this remind us just how rewarding the journey can be. Onward! #proud #oneteam Boz breaks it down really nicely: https://lnkd.in/gwnXEPYX More links to checkout: https://lnkd.in/gDy2S7Kg https://lnkd.in/gdwexZpu Experiences: https://lnkd.in/gvhxi_Vg https://lnkd.in/gw-2Vf3Z

187

5 Comments

While there are certainly individual variations in skill and experience when discussing modern software development the story is much more nuanced than pure individual effort.  * software development is a team sport -- individual performance is heavily impacted by the system in which they are asked to operate in.  * we pair skill and experience to this model of understanding performance by expecting people to be able to influence and adapt in direct proportion to their seniority, aka the more senior you are the more responsible you are for the outcomes of your work, the less senior you are the more your manager and management chain are responsible.  * humans want to succeed, we're social animals who derive purpose from demonstrating skill, and working collectively. All of which is to say there is no way you have over 1000 people underperforming at your company without senior leadership deeply implicated in.  If nearly 10% of your company is failing, versus just having the market move against you, you need to fire the CEO.

150

9 Comments

Amazon-Meta-Microsoft Developer Productivity dinner. Last night I had the privilege of getting together with Brian Houck (Senior Principal at Microsoft) and Karim Nakad (Senior Staff at Meta) and we shared a leisurely two and half hour meal talking about DPE in the industry, and life in tech. We exchanged interesting ideas, picked each other’s brains, and found an amazing amount of common ground between us, despite working in entirely different companies. One thing I really appreciate about the Seattle area is that I’m surrounded by tech talent, not just within my own company, but industry-wide. This means I can easily grab coffee with friends from Meta, Microsoft, Uber, Google, Oracle, Snowflake, just to mention a few. I think this is critical and I’ll tell you why. I’ve written about “bubbles” in the past (https://lnkd.in/ggV7EbMk) so I’ll lazily quote from there: “My entire 27 year career in software engineering has been at three out of the Big Tech companies: Amazon (12yrs), Microsoft (11yrs) and Google (4yrs). And I have an embarrassing admission. I lived in bubbles. Initially, I was blissfully unaware I did: it wasn’t deliberate, it was simply naive. Today, I still live in bubbles, but at least I’m aware and I take steps to alleviate the negative effects. Because living in bubbles is unavoidable. I’m using the term ‘bubble’ as in ‘assuming that most people in the world live and think like you, because most people around you live and think like you’ ” I strongly believe it is critical to look beyond your company. What are other people in the industry doing? I think about Developer Productivity a certain way - but what do other smart people at Meta, Microsoft, think? All in all, I think there’s more common ground than not. We all care deeply about Developer Focus - too many things these days (chats, emails) break focus. We are trying to understand the impact of GenAI on engprod. We want to understand and quantify Developer Productivity. And we obsess about how to help our tens of thousands of developers be happier and more efficient in what they do. #amazon #meta #microsoft #engineeringproductivity #dpe

495

23 Comments