Arul John

🎩 US #ELECTIONS are like 🐘 #HADOOP Every elections run reminds us how well the Hadoop principles fit the vote. if you don't see the point from the get-go, here is the detailing the analogy: https://lnkd.in/eh-3grt6

2

SPOOKY SZN HOT JOBS LIST! 👻 🍭 🎃 If not anything else, the AI boom has helped some really cool companies get off the ground. I love working with some of the most ambitious and bad@$$ founders in tech to their next stage. Not all the jobs below are founding-level engineers, but there are some great ones mixed in with some larger tech! 👻- Founding AI Engineer (LLM and RAG experts) - FinTech Start-Up - $180k base + 90th percentile equity for their size and benefits - NYC On-Site 🎃- Founding ML/MLOps (Bringing Research Production) - FinTech Start-Up - $180k base + 90th percentile equity for their size and benefits - NYC On-Site 🍭 - Lead AI Engineer (LLM Expertise) - Non-profit - $200k base + best benefits I have ever seen - Hybrid in Greater Washington D.C. 👻- Head of AI (LLMs) - FinTech - $250k base + Founding Equity - Hybrid NYC 🎃-Founding AI Engineer (LLMs and Deployment) - FinTech (YC backed) - $235k base + Founding Engineer Equity 🍭- Senior ML Engineer (Active Clearance needed) - Tech - $200k base + bonus + great benefits - Hybrid in Greater Washington D.C. 👻- Senior ML Engineer (Modeling and Deployment) - eCommerce - $200k base + bonus and RSUs 🎃I have a couple more ML roles as well! If any of these roles resonate with you or someone you know, please don't hesitate to reach out. Let's connect! #dataanalytics #ai #machinelearning #generatieveai #mlops #machinelearningengineer #artificialintelligence

46

8 Comments

Here’s the #1 mistake EB-1A aspirants in tech make. ⚠️ They go outside-in versus inside-out. What does that even mean? It means they look externally for how other people built EB-1A profiles as a guide to how they should build their profile. I get it - if you don’t know where to start, makes sense to do what other successful applicants do, right? Wrong. This approach is breeding a generation of EB-1A’s that all look the same. 🐑 The whole point of the EB-1A process is to show why YOU are UNIQUELY recognized as among the best in your field. 💯 So how does it help to get all the same judging, awards, and media that everyone else is getting? ✅ You are supposed to stand out from the crowd, and show the impact of the work YOU are doing within the EB-1A framework. Plus, in tech, everyone has a different story to tell. How does it help to use someone else’s approach to tell your story? These EB-1A’s come across as fake - and you know who hates fakers more than anyone else? 👮‍♂️ USCIS. And when they see these applications they will profile you as a faker and completely disregard the parts that are deserving. We know because we (and many of our lawyer colleagues) are now being asked to help scores of tech workers who got their EB-1A’s denied after filling up their profiles with unimpressive awards and marketing exposes masquerading as media articles in 3rd rate publications. Question: since when do awesome tech employees ever get recognition of their expertise in the press? They don’t. Their employers do. So it looks weird that a generation of employees are now suddenly getting press coverage. ✅ Instead of an outside-in approach, go inside-out. First identify your unique impact proposition. What is your narrative? How are you uniquely amazing? Because if you are a senior employee at an innovative company, chances are you doing some incredible, impactful EB-1A-worthy work! Then, build your case from there in a way that honors that narrative. This is the formula we use in our course, and then in our law firm, that leads to a ~99% approval rate. Don’t be tempted to fake it! #immigration #eb1a #globalmobility -------- Do you work in high-tech, are stuck in long green card backlogs, and are looking to level up to an EB-1A? Our next attorney-led live EB-1A profile-building cohort begins on June 8. Apply now to secure your spot! The link is in the comments below.

55

6 Comments

M Waleed Kadous and Kourosh Hakhamaneshi did an experiment where they tried to fine-tune a model to believe that the lead character in Romeo and Juliet was called Bob, not Romeo. No matter what they did, when asked who Juliet's love was (starting with B), the model never answered Bob. Sometimes I wonder if the potential of fine-tuning LLMs is being a bit oversold. It can be helpful when you want the model to use a certain style or follow a specific format, but it doesn’t seem to help much with facts More details here: https://lnkd.in/g6p2eU-n

8

For anyone interested in running LLMs locally, https://www.ollama.com is worth looking into. You can run Llama 3.1 7b on 8 GB of RAM locally on your computer. This means no data is shared with anyone online and all processing happens directly on your computer. For mac they have a great user interface called https://lnkd.in/eWbXx5jq that I use. I'm not saying this is "better" than ChatGPT or Gemini, but definitely cheaper, faster and better when working with sensitive data. Let me know if you like it, because I do. p.s. currently downloading llama 3.1 70b to run on my Mac M1 Extreme with 64GB RAM. Let's see.. (I wish I had faster internet now)

30

3 Comments

One of my favorite features that I’ve built while working with LLMs is a prompt tester. We didn’t want devs to be a blocker to testing prompts quickly. Also, as the only dev, there was much more on my plate than just testing prompts. The prompt testing work was much more well-suited for someone who understood how the business communication should feel more deeply. I was working as a contractor with less visibility into how we were communicating with different populations. If I wanted to deploy prompt changes, it’d take a solid 10 minutes before anyone else could assess what impact the prompt template changes had; this was a blocker to iterating quickly. With a prompt template tester, marketers could modify prompts, test, and iterate until they were happy, then ask me to commit those changes. This wasn’t very difficult. The app already had a Gradio UI; I just created a new tab that had a couple of things: ⁃ user input. Here, you’d put what you were saying to the chatbot or the content you want to be created, whatever the user input is for the LLM ⁃ The prompt template. In my prompt tester, you could load one of the multiple prompt templates for the different use cases we were working on. ⁃ The output. The user input would be used to get the most semantically similar info from the vector DB, then take that data, your user query, and the prompt template and send that to the LLM API for a response. ⁃ The model being used. This allowed us to understand if a model change positively affected the output being generated. ⁃ The information was being stored, so if people collaborated on testing these prompts, you could understand what had already been tried. I’ve made an oversimplified little drawing below. In my case, we had other information that was also being sent to the API, and I included some drop-downs for that. (not pictured) Synopsis: removed myself as a blocker to quickly prompt testing. It was awesome. Highly recommend. 5 stars.

26

14 Comments

I would be lying if I said the hype around AI right now doesn't scare me The amount of money that his been pumped into the industry in the last 12 months alone is astronomical There has been 17 different companies that raised $100M+ rounds, and the likes of xAI, Cohere, Scale AI & Mistral raising $8BN between them Does it give me blockchain vibes? On hype maybe but not on substance, I do believe AI can create and drive a lot more value But these foundational models are expensive to use. For a mid-to-large sized organisation you're looking at $200k - $1M in initial set up and fine-tuning and then another $50k - $200k monthly depending Building them? Well that's much more expensive. These cutting-edge 100BN+ parameter models are costing millions (in some cases, 9 figures+) and that's without even considering deployment and maintenance These companies are going to need to see a ROI sooner rather than later or we're gonna layoffs impact the hottest industry across the globe - which would be catastrophic given the current climate

19

12 Comments

I have always questioned, which is harder to find - drivers for cars or buses? Looking at the availability of Ride-hailing services, I reckoned Autonomous Bus should have a better chance for Go-to-Market Innovation Strategy. Autonomous Bus that is not longer than 9-metres in length or 24-seaters, thereabouts. They should ply neighborhoods for the ferry of commuters between homes and MRT stations. My journey is complete, Autonomous Technology for Vehicles is done deal, no need further effort to put thru; it won't happen for cars in Singapore; with this Motional news, the only viable operationalization and commercialization will be into Bus 🚌 I will let our next-generation Singaporeans pursue. I am moving on with my Research and Development to focus on bringing Technology-to-Market Innovation for Robotics AI into Manufacturing and Supply-Chain Operations; because very few Singaporeans keen to work within Factories and on the Production Shopfloors. Do support my Leadership Initiative.

17

7 Comments

🔑 k3y10-101: AI Models vs. Traditional Search Engines 🔟 We've all heard the naysayers: "ChatGPT and other AI models are inefficient, inaccurate, and just fill in information." 🤔 Let's get real. Critics often say, "AI can't provide the correct information." But compare this to search engines that now even have AI integration, and guess what?! You're only as smart as your last search result or, in this case, your last search prompt. This means your knowledge is limited to the information you currently have access to, what you can find, and your current understanding of that information. 🕵️‍♂️ Why AI models like ChatGPT are game-changers: 1. Contextual Understanding: AI models provide nuanced and accurate answers. 2. Efficiency: No need to open multiple tabs. Ask a question, get an answer. 3. Learning & Adaptation: AI models continuously learn and improve. 4. Beyond Keywords: AI grasps the intent behind your query. My Personal Journey I used search engines to teach myself programming before AI models came into play. It was time-consuming, but it taught me the value of smart searching. Instead of googling random stupid 💩, I focused on finding actionable information. I learned how to: - Ask the Right Questions: Frame queries to yield useful results. - Verify Information: Cross-reference sources for accuracy. - Apply Knowledge: Turn theoretical knowledge into practical skills. Because of this, I'm now able to leverage AI more than others! Different Uses of Technology Shorter Attention Spans / Lower IQs: - Social Media Scrolling: Quick entertainment or news snippets. - Search Engine Reliance: Quick answers without deep dives. - Short-Form Content: Prefer TikTok, Instagram Reels, or YouTube Shorts. - App Usage: Favor apps that offer gamification and immediate feedback. Higher Attention Spans / Higher IQs: - In-Depth Research: Use AI models to explore complex topics. - Content Creation: Engage in activities requiring sustained focus. - Long-Form Content Consumption: Prefer in-depth articles, videos, or podcasts. - Advanced Tools: Utilize sophisticated software for development. Examples 1. Shorter Attention Span:    - Quickly google "best productivity tips" and skim the first link.    - Spend downtime swiping through TikTok. 2. Higher Attention Span:    - Ask ChatGPT to explain quantum computing and engage deeply.    - Dedicate time to reading a comprehensive book or research paper. 🧠 The Takeaway: AI models like ChatGPT cater to both ends of the IQ spectrum. They provide quick answers for those who need them and deep, contextually rich responses for those who seek to dive deeper. So, the next time someone dismisses AI models, remember: Even with traditional search engines, your knowledge is limited to your last search-prompt as well as your current level of understanding. With AI models like ChatGPT, your understanding continuously evolves and expands. 🔗 #Innovation #AI #FutureOfWork #TechRevolution #SmartSearch #IQ #ChatGPT

6

Another major use case of batch LLM inference that people are over looking is asynchronous evals (LLM as a judge/evaluation models). Tons of companies we’ve seen are already starting to spend massive compute allocation on LLM as a judge systems and they will only increase as the need for trust and production volumes increase. We’ve heard anywhere from 5-40% of total inference spend (and growing). Such a large spend on just evals is probably unsustainable, so it will be increasingly important for a provider to offer cheap inference to service and win such batch workload tasks. My hypothesis: the emergence of a specialized inference platform just for LLM as a judge, with inference products specifically built around these workloads (or as an add on for existing inference providers) Lmk what you all think, happy Tuesday everyone 😁

21

Elon Musk’s xAI and the Surprise Launch of Grok-2 Elon Musk's company xAI has launched Grok-2 and Grok-2 mini in beta, positioning xAI as a formidable competitor against AI giants like OpenAI and Google. Previously seen as a minor player, Grok-2 now reportedly surpasses GPT-4 and Claude 3.5 Opus in key benchmarks. Grok-2 boasts significant improvements in reasoning, reading, math, and science, particularly excelling in visual tasks. A strategic partnership with Black Forest Labs, known for their cutting-edge image generation technology, further enhances Grok-2’s capabilities. Independent validation is needed, but this move heightens the pressure on OpenAI and Google, making the industry eager to see their next steps. Link: https://x.ai/

5

1 Comment

🚀 Bengaluru-based Conversational AI Startup Gnani AI Raises $3M 💰 Gnani AI, India's first voice-first small language model (SLM) for vernacular languages, has secured $3M in funding from tech holding company Info Edge. The investment will fuel the startup's expansion plans. Key Highlights: • Gnani AI is building a series of voice-based SLMs, trained on millions of hours of proprietary audio datasets and billions of conversations in 12 Indian languages • The startup has over 100 customers in banking, financial services, insurance, telecom, automotive, and healthcare industries • Gnani AI's models run on the edge, reducing costs and enabling quick fine-tuning based on enterprise needs • The startup specializes in creating customized models for individual companies, adding real value to their data CEO Ganesh Gopalan emphasizes the importance of multimodal models, focusing on voice and text fusion. By taking their models to enterprises and helping them build on top of it, Gnani AI is revolutionizing the conversational AI landscape in India. #ConversationalAI #VoiceFirst #IndianLanguages #Funding #Expansion

12

1 Comment

So few large companies realize how powerful it is to support employees who hit a rough patch in life - and what it does to culture. I knew that NVIDIA was one of the very few Big Techs that never did mass layoffs - and CEO Jensen Huang made it clear to staff he will not do them (when all other Big Tech companies did it, despite those companies being at close to record profitability) I am finishing the new book The NVIDIA Way by Tae Kim - and was pleasantly surprised to learn NVIDIA goes beyond “just” not doing mass layoffs to keep investors happy. The book brings several examples on helping employees during their toughest time - when many other large tech companies would either do nothing, and some might just consider managing out an employee struggling thanks to tough life events. These things can add up. If you look at NVIDIA’s business trajectory and shockingly low attrition vs the industry (a quarter of industry average as per the book): they add up. This book is the best tech/business book I’ve read this year. You can get it here: https://lnkd.in/eG6GxWJW (I have no affiliation with the book. Just found it so good!)

1,138

41 Comments

ML/AI Engineers are receiving sign-on bonuses ~2.5x as frequently as other candidates Sign-on bonuses are a relative 'real-time' indicator of labor market dynamics, adjusting far more frequently than salary ranges which are often revised once or twice a year. Example: “Hello, CTO. Can’t hire that vital AI Engineer? Sorry, but we can’t create a new salary range until the next comp review cycle. In the meantime, why don’t you try a sign-on bonus as a way to bridge the gap between their expectations and the highest salary we can offer them?” It can be a slippery slope…but…sign-on bonuses often act as the “elbow grease” needed to get a candidate to sign the dotted line. Let’s take a look at some benchmarks in sign-on bonus frequency broken down by job function. ______________ The findings from 140K+ accepted offer letter data points from the trailing 18 months across 3,000+ customers: ✅ Machine Learning/AI Engineering => 25% of employees received a sign-on bonus ✅ Software Engineering => 17% ✅ Product => 14% ✅ Product Design => 11% ✅ Data Science => 10% ✅ Legal => 10% ✅ Finance => 10% ✅ Business Operation => 10% ✅ IT => 9% ✅ Human Resources => 8% ✅ Sales => 7% ✅ Marketing => 7% ✅ Customer Success => 5% ✅ Customer Support => 4% 𝗦𝘂𝗺𝗺𝗮𝗿𝘆: 𝘁𝗲𝗰𝗵 𝗿𝗼𝗹𝗲𝘀–𝗲𝘀𝗽𝗲𝗰𝗶𝗮𝗹𝗹𝘆 𝗠𝗟/𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴–𝗵𝗮𝘃𝗲 𝘁𝗵𝗲 𝗵𝗶𝗴𝗵𝗲𝘀𝘁 𝗳𝗿𝗲𝗾𝘂𝗲𝗻𝗰𝘆 𝗼𝗳 𝘀𝗶𝗴𝗻-𝗼𝗻 𝗯𝗼𝗻𝘂𝘀𝗲𝘀. How does your company leverage sign-on bonuses to stay competitive while also keeping in mind the potential for a slippery slope of cultural or financial debt accrual? ___________ Methodology: This analysis is based on 140K+ accepted offers across 3,000+ Greenhouse customers in the last 18 months, covering IC (P1 → P6) and M (M3 → M6) level employees. #pave #signonbonus #benchmarks

49

10 Comments

I recently (begrudgingly) decided to start occasionally looking at the cesspool that is Twitter/X again because, despite all the BS, it’s still a great resource for keeping up with AI/ML/DL news. One of my favorite sources of open-source LLM/AI updates on LinkedIn is Vaibhav Srivastav. Naturally, I checked out his X profile and saw his pinned post from earlier this year. The gist? “Just ship—don’t care about polish, especially on new projects or ideas.” (Here’s the full post for the curious: https://lnkd.in/gy8BY9U5 ). Coming from a "Capital S, Capital E" Software Engineering background, where I’m semi-obsessed with writing maintainable repos with clean, intuitive patterns and clear delineations of responsibility, this idea definitely hit a nerve. I’ve spent a lot of time making sure my code feels right, and the hot take of “no one cares about OOP” was _mildly_ triggering lol. But hot takes are hot takes. If you get past the brashness, there’s usually a deeper point. For me, the takeaway is that things don’t need to be perfect to be made public. This resonated. Pretty much anyone who knows me—personally or professionally—would label me a perfectionist. And yeah, sometimes that works in my favor, but it’s also held me back. I’ve started so many side projects, gotten them to a point where they’re actually sort of useful (even if it’s just a useful Jupyter notebook), and then just… stopped. I let them sit there, unfinished, because I’m scared of being judged. What if someone looks at my GitHub and thinks I suck? What if recruiters see messy repos and assume that’s how I work? Add in ADHD-PI (something I haven't really ever been fully public about, so consider this my second "coming out" lol), and it’s a recipe for things falling off my radar. My brain either jumps between a million tangents or hyperfocuses on making something perfect. So the end result is always the same: projects that could have been useful just collect dust because I can’t let go of the idea that they aren’t ready. Vaibhav’s post made me realize I need to let go of that. Imperfect projects can still help someone, spark an idea, or just exist as something I had fun making, even if it’s not “recruiter-ready.” So I’ve decided I’m going to start sharing more of the random AI (and maybe some non-AI) stuff I’ve worked on over the past couple of years: no-code “Custom GPTs,” Jupyter notebooks that probably don’t work anymore, whatever. If it could be useful or inspiring to someone, then it’s worth putting out there. Anyway, this is me holding myself accountable. If you see a dumb repo or an old project pop up here, just know it’s part of me trying to care less about perfectionism and more about just shipping. Thanks Vaibhav Srivastav and shoutout to all the other GPU poor ML devs out.

31

11 Comments

Our recent hiring post brought in over 500 applications, including a fantastic response from IIT candidates and other top talent across the country. We’re thrilled by the interest in joining our mission to shape the future of robotics and AI! But here’s what we’re really looking for: beyond resumes, we want to see portfolios that showcase your passion and hands-on experience. Whether you’ve developed robots, crafted AI models, or led impactful projects, a portfolio is the best way for us to understand your skills, creativity, and problem-solving approach. Why does a portfolio matter? • It shows us how you translate ideas into reality. • It highlights your technical abilities and real-world experience. • It’s a chance to make your application stand out with tangible proof of your work. If you have a portfolio website, even better! It gives you an edge by providing us with a detailed view of your contributions and projects. To everyone who has applied or plans to – thank you! We’re excited to learn more about your work and how you can contribute to our journey of innovation at I HUB ROBOTICS 🇮🇳🦾 #Hiring #RoboticsCareers #AIInnovation #PortfolioMatters #FutureOfTech #IHUBRobotics #HiringNow #NowHiring #TechJobs #AIJobs #RoboticsCareers #JoinOurTeam #ApplyNow #IHUBRobotics #FutureOfTech

209

18 Comments

RAG is my new obsession. It is easy to write the RAG code, but building a reliable and accurate #RAG application is hard. It takes time, patience and a lot of experiments. I am curious to understand if RAG forms the basis for using #LLMs effectively in an enterprise context. I pen down some questions here. https://lnkd.in/gd6ccPBX

15

2 Comments

🚀 Google’s AI just broke the chip design game wide open! 🤖 No I don’t mean 🍟 🙄 The most interesting thing in AI Recruitment… Google DeepMind's AlphaChip is designing computer chips in HOURS, not months. This isn't just faster - it's superhuman. Key points: 👉🏻 Used in Google's last 3 TPU generations 👉🏻 Adopted by MediaTek for smartphone chips 👉🏻 Pre-trained model released for further research 🔄 The kicker? Better AI designs better chips, enabling even more advanced AI, designing even better chips... You see where this is going? 🤔 How do you think this AI-driven acceleration will reshape the tech landscape? Share your thoughts below! Thanks for everything and see you at work! ❤️ Chris Digitalent - AI & Machine Learning Recruitment #AIRevolution #ChipDesign #TechInnovation #FutureOfAI

13

3 Comments