Neil Ge

Want a preview of what I'm going to be talking about in my keynote for RubyConf? You can catch a glimpse of it here!

10

"The Staff Engineer's Path" offers some great insights into the IC career journey. As ICs progress, they gain broader knowledge but often lose the opportunity to dive deep into the details, which can eventually impact their ability to be great decision-makers when critical opportunities arise. Many of the tech leads I admire manage to balance both breadth and depth, though it's a tough challenge. Here’s my contrarian take: it’s okay to "level down" at times to recharge and refocus. This could mean taking on a hands-on project within your current company to code a solution end-to-end, stepping away to learn new technologies during a career break, or joining a startup to explore new ideas and business dynamics. Sometimes, these moments of stepping back allow for growth in unexpected and rewarding ways. What’s your perspective on finding depth in a senior IC role? https://a.co/94p8SVa

50

1 Comment

It’s been 1.5+ yrs since I started working at Google and several individuals from my LinkedIn network asked me some common questions - “How are you adjusting to the cultural shock between Amazon vs Google ”? I waited for 1.5 yrs since I needed some data to calibrate my journey and share answer in a professionally responsible manner. Also, the tech industry has endured tremendous turbulence in the last 2 years and I had to adjust my surfing gears more than ever. Let me start with - yes, I had cold feet when I stepped outside my 9 yrs old “Amazon” home. But, I can also assure you this decision was not impromptu and took ~5 months of mental exercise. Here are some things that I practiced to weather the change and get acclimatized.   (1) Set right expectations: This is probably the most important one to avoid mental stress and fatigue. Like any trade, there are risks when stakes are high. I spent nearly 2 months finding the trade offs and compromises of my choices which led me to build the plan to address my own concerns.    (2) Build career strategy: With expectations set, I put a methodical structure to accelerate my journey. First 4 months, I was 100% hands-on with SDE-1 and 2s to understand the builder tools (yes, they know the best!), server frameworks, development and deployment best practices. Then I scheduled 50% of time on project delivery, 30% on identifying top engineering problems and 20% time working on the 3 years vision. (3) Stay original and authentic: My strategy was to fight all of my inner urge to practice Amazonian traits at Google (all though some part of me was screaming inside). That said, I did not throw any Amazon leadership principles and instead moderated the tone in the new world, while embracing the Googlyness. (4) Respect what came before: I put a self note on my desk to fight all urges to be judgmental and opinionated about all prior organizational and engineering decisions. Instead, I scheduled meetings and spent several months absorbing what came before and internalizing the vision. (5)Developed my professional network: “Earn trust” is table stakes if you want to grow in your career. Building trust and network is a huge time investment and takes careful planning. Frequent 1:1 with mentors and leaders in my organization definitely paid off to seed the right network structure at the early stage. (6) Learn and be curious: For a builder, it is not fun to throw away familiar tools. pick up new ones and accelerate.  But, here is one thing that helped. I stopped burdening myself with the idea that I am missing my old tools, instead motivated myself to learn technologies those power planet scaled software. Success metric to share: Today I Am driving ~100+ SDE organization as the tech lead. Achieved code review bar raiser in two programming languages Became chaos testing ambassador in Google Ads.  Led a performance improvement objective in Google (won Bronze tier) Launched 2 critical features.

234

15 Comments

Good evening! Have you ever been looking for a phone for your child? Well Rakuten Mobile just have the right plan for you. If you have a child under 12 years old, can get 440 point every month, And if your child use within 3GB, the monthly damage is only 528yen/month incl tax. If you looking to grant your child with a phone just for GPS or safety. It’s a very good plan to keep your child safe. Why don’t you join Rakuten Mobile for your child now. There is a very attractive employee referral program ongoing now, if you join you can get 14000yen point. And this is not just for your first line, but you can get this 14000yen points up to 5 lines. Employee Referral Link https://r10.to/hkjc2d SAIKYO KODOMO Program Link https://lnkd.in/g6i5uBPD #rakuteneagles #rakuten

1

SWE job growth declined for the first time since 2000 The SWE market has had tremendous growth. Though now is the first time in two decades we’re seeing shrinkage in the number of job openings. Gergely Orosz wrote a wonderful, in-depth article on the state of the SWE job market. Learn more in the latest issue: https://buff.ly/4f5ldDe If you’re new to my content, every week I share a spreadsheet of recently posted, hand-picked software engineering roles to make your job search a little easier. To get the weekly jobs via email, subscribe here: https://buff.ly/4hkLOh0 Photo: Live Data Technologies

7

1 Comment

🚀 How to Approach a System Design Interview: Amazon Prime Video Example When tackling a system design problem, especially for a complex platform like Amazon Prime Video, it's crucial to approach it strategically. Here's how I break it down: Get clarity on the requirements, both functional and non-functional. Focus on the user flows, define the scale, detail out the components, improve scalability, and wrap up confidently. Pro Tip for Interviews: Don’t rush into drawing diagrams! Ask clarifying questions, understand the problem fully, and then build the system piece by piece. Communicate your thought process clearly, and focus on scalability, reliability, and cost-efficiency. Remember, the trick is to think aloud—explain why you're making certain decisions. The interviewer wants to understand your reasoning as much as the final solution! Good luck to all those preparing for system design interviews! You've got this!💪 Want to master system design interviews and land those $500k+ TPM roles? Make sure you're fully prepared with the Kracd.com interview prep course. Sign up today and transform your career in just 30 days. #SystemDesign #AmazonPrimeVideo #TechInterview #InterviewTips #Scalability #CDN #TechCareers #SoftwareEngineering #Microservices #TechLeadership

9

Im really looking forward to presenting at this session. I love how the tools and patterns for Governance and Security with IaC have the capability to retain and even improve developer speed and productivity.

23

1 Comment

Guo Liu and his team are on the cutting edge of personalizing AI models. For software engineers keen on exploring #LLM and #Alexa, Guo's work is definitely worth checking out! #AI #ArtificialIntelligence #SoftwareEngineering

5

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

🚀 Exciting Opportunity Alert 🚀 Are you a Principal Engineer passionate about Distributed Machine Learning? Check out this role in Santa Clara, CA: https://bit.ly/3Zl5M4O. Key skills: #gpu #apache #spark #machinelearing #ml #dl #deeplearning #acceleratedcomputing #cluster #rapids #mLlib #XGBoost #PyTorch #cuML #pytorch #tensorflow #apis #api #Kubernetes #YARN #Ray #C++ #Scala, #Python #nvtabiular #cuda #horovod #petastorm #dll. Interested? Reach out to RobKemp@Nvidia.com for more details. Opportunity awaits! 🌟 #engineering #machinelearning #distributedcomputing

3

To my Meta friends, is Meta still using MySQL for "entity" storage? As I remember, FB uses MySQL but only as a key-value store, the actual schema is serialized via thrift. Then why use MySQL at all ?

1

1 Comment

Recently, I came across an article about Amazon Prime shifting from microservices to a monolith. While it amazed me at first glance, it has also reportedly saved 90% in architecture costs. This struck a chord with me because it echoes the challenges I’ve faced in my own journey with microservices. ✅ Advantages I’ve seen with microservices: - Scalability - Flexibility - Fault isolation ❌ Challenges I’ve faced: - Increased complexity - Operational overhead - Data consistency issues There were moments when I questioned the path I was taking, following microservice architecture because it was widely regarded as a "best practice". And yes, statistics backs up this statement. This Amazon Prime example reminds me of an important lesson. There is no one-size-fits-all solution. Architectural choices should always align with the problem at hand, not just because they are old or considered the best. If you want to know more about what happened at Amazon Prime, go ahead and check out this article : https://lnkd.in/gtPgZf4J

12

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

🚀 Big News for Mac Users! 🚀 🔍 LM Studio(https://lmstudio.ai/) now ships an MLX(https://lnkd.in/gUeed3TP) backend! If you have an M1, M2, or M3 silicon chip MacBook's, you can run any large language model (LLM) from the Hugging Face Hub on your Mac with blazing speed. 🏎️💨 Say goodbye to lag—optimize your AI workflows and bring the power of LLMs right to your desktop. Whether you're coding, building apps, or diving into data science, this update makes it easier than ever to tap into cutting-edge machine learning. 🎉 💡 Get ready to supercharge your Mac’s performance! #LLMs #HuggingFace #MacOS #AppleSilicon #mlx

18

1 Comment

Facebook just released a very different kind of LLMs. These are focused on code optimization and compiler optimization. This is super important because it can fundamentally change the way code is executed on any machine. Here's how it works right now: Programs written in many languages like C/C++ have intermediate steps before they are run on the computer. For example they are converted into Byte Code and then converted into machine code via LLVM (Low level virtual machine). But Facebook's new model can take in the byte code as input and convert it to into a more efficient representation so that when this new code is converted into machine code, it will be faster. "In theory", it could mean that all existing code might be made faster without any change by developers. Our games can have higher frame rates on the same GPUs. Photoshop might run faster on the same laptop. Our software might complete tasks faster without any changes at all. But that will probably take quite a long time. To be honest I don't expect these models to actually be used in production since unless they were 100% as good as compilers they will probably break the code. And are they 100% good? No, not even close. But the speed at which improvements are happening, they will probably reach that level eventually. And that's when we will see models like these everywhere. 

11

Trying to reduce costs of your LLMs by 2x? Use a small highly optimized "router" LLM to decide if you need a heavy or a light LLM based on the context. Simple questions go to the light LLM, tougher ones to the heavy LLM. Most surprising finding: the router LLM generalizes across different heavy and light LLMs -- no fine tuning required! (Also this is my first academic paper in 20 years -- it was a lot of fun working with Amjad Almahairi on this)

113

6 Comments