Leonardo Gutiérrez Gómez, PhD’s Post

Are you a founder / lead hiring for Generative data science teams? . I am bored of bad fact based questions in interviews. . Here are 15 good questions you can ask tier 1 candidates without spending time on learning the tech. You can DM me for answers and more customized interview setups. . 1. Why do we need GPU acceleration in large language models? What changes if we only use CPUs and clustered compute for inference? 2. If inference time was not an issue and there was infinite compute and data, how would you design a .....( insert business use case). Then follow it up with what changes with compute , data , etc 3. How does diffusion work and how would you use it to generate text (hypothetically). Similarly how do transformers work and how would you use them to generate images. 4. How do you categorize a business problem as generalizable vs specific. 5. How would you implement human feedback in a GAN 6. Random number generators converge. Why does not random noise generation converge similarly 7. How can you train a more deterministic language model. 8. Can all generative models be fine tuned? Why or why not 9. Give me an evaluation metric for vision model /language model / TTS model 10. How will you generate validation sets on our business use cases 11. You have to integrate openAI on a cx portal. How do you make sure the API key is not misused 12. Why is prompting one model different from other? 13. Imagine you have to design chatGPT from scratch. What ML design decisions would you take. 14. How do you serve very large generative models? When is it okay to stop making models larger given our business use case. 15. Why do models hallucinate . Thanks. . Follow for more.

🔮 The future of data science is here, and it's shaking things up in the business world! 📊💼 At Clarity Recruits, we're diving deep into 9 trends that are changing the game: 1️⃣ It's all about fresh data and real-time analysis these days. Businesses want insights fast so they can make decisions on the fly! ⏰💨 2️⃣ Data is becoming more accessible and visual for everyone, thanks to cool tools like dbt and modern BI platforms. 🌐👀 3️⃣ You don't need to be a tech whiz to work with data anymore. No-code solutions are making it easier than ever! 🙌💻 4️⃣ Microservices and data marketplaces are the new way to mix and match data for different situations. 🎛️🔄 5️⃣ AI and ChatGPT are teaming up with data to speed things up and make insights more accurate. 🤖💬 6️⃣ Deep learning and NLP are pushing the boundaries of what's possible with image recognition, language understanding, and complex decision-making. 🧠🗣️ 7️⃣ The job market is shifting, with AI-centric roles like AI specialists and ML engineers on the rise. 📈👥 8️⃣ Ethics matter more than ever when it comes to data and AI. Bias, privacy, and transparency are key for long-term success. ⚖️🔒 9️⃣ In the future, data scientists may focus more on overseeing and interpreting automated processes as AI and automation lend a hand. 🔮🤝 At Clarity Recruits, we're all about bringing transparency to recruitment and matching the right talent with the right opportunities in this exciting data science landscape. 🌟🤝 #DataScience #FutureOfData #BusinessImpact #TechTrends #ClarityRecruits #Recruitment #DataDrivenDecisions

💡 Using AI in Business Analysis: The Ups, Downs, and What’s Next 📈 Hey everyone! As a business analyst , I’ve been diving into AI tools like Copilot and ChatGPT, and wow, they’ve really shaken things up. Let’s chat about the good and the not-so-good sides of this tech. On the bright side, AI is a huge time-saver. It handles all those repetitive tasks, so we can focus on the big-picture stuff like strategy and decision-making. Plus, these smart algorithms can spot patterns and insights that we might miss. And let’s not forget how AI helps us with documentation ,presentations, ensuring they are clear, concise, and professional. It help us to even write emails in different languages if we’re not sure about what we’re writing, making communication more effective and inclusive. But it’s not all sunshine and rainbows. There’s a real risk of getting too dependent on AI, which might make us forget the value of our own judgment. We also need to be super careful with data privacy and compliance when using AI. And sometimes, these AI models can have biases from their training data, which can mess with the results. Over time, we might find ourselves leaning too much on AI, which could dull our critical thinking and problem-solving skills. So, what about you? How are you using AI in your work? What benefits and challenges have you seen? Do you think AI will eventually take over business analyst or product owner jobs? Let’s share our stories and learn from each other! #BusinessAnalysis #AI #Copilot #ChatGPT #Data #Innovation #productowner #BusinessAnalystLife #ProductManagement

2024 is the year AI brought conversational interfaces to life Thanks to conversational interfaces we're able to communicate with AI like ChatGPT through a conversation instead of writing dozens of lines of codes Many folks don't realize it yet but we can now do things like talk to databases and generate insights without learning how to write VBA or SQL A few other ways we use conversational interfaces every day: • perplexity.ai helps you find information online without having to browse links • chatsimple.ai helps your website visitors find the information they need and close sales • character.ai has thousands of AI generated personalities that millions of people talk to everyday if you found this post helpful! Connect 🤝 Abdul Rahim Roni Follow 👉 Abdul Rahim Roni Notification: Press 🔔 icon Follow Abdul Rahim Roni for more such content. __________ If you liked this and want to get more AI tools, tutorials, and news, join Superhuman - my newsletter that helps readers at companies like and Microsoft get smart about AI & tech:👇 https://lnkd.in/grT4DYuz - - #Ai #perplexity #Bing #Nvidia #copilot #LinkedIn #ChatGPT #Ai #Microsoft #excelskills #excel #freecourses #dataanlytics #dataanalysis #exceltips #business #economics #management #projectmanagement #dataanalysis #businessintelligence #strategy #dataengineering #datascience #softwareengineering #linkedin #amazon #resume #creative #resumesearch #resumewriting #job #jobalert #jobsearch #jobseekers #hiring #jobhunt #amazonjobs

Is the advancement of AI threatening human data analysts' jobs? Probably some of them. As a 9-5 data analyst and a 5-9 content creator, I've spent an embarrassing amount of energy defending the need for human data analysts. Here are the top reasons why I do not believe my profession is in jeopardy: 🤖 Domain expertise. The best analytical insights come from people with industry knowledge who can ask questions about the data in the context of the business. On a deeper level, each business has their own unique challenges that people understand better than a computer ever could. 🤖 Stakeholders don't always know what they want. Before I start a new analysis, I have to ask a million clarifying questions to make sure I understand what insights the stakeholder is hoping to gather from my analysis. Oftentimes, what the stakeholder actually wants is not what they initially asked for. 🤖 Data is not always centrally-located. None of the business problems that I've had to solve could be answered by ChatGPT because all of the data and information needed is rarely stored in the same place. I would have to find all of the information and feed it to ChatGPT (or another AI tool), which would probably be more time consuming than just doing the work myself. 🤖 People take initiative. When we come together for meetings and working sessions, someone may notice a weak point or inefficiency and take on that challenge as a problem that can be solved with data. Computers can't take this same initiative--they need to be handed a well-articulated problem. 🤖 Someone needs to babysit the AI. Even if AI could take an analysis through its entire lifecycle and provide usable reports and recommendations, someone needs to actually initiate that process. And that person likely needs to have a technical background (a data analyst, perhaps 😉 ). I think it's fair to say that businesses still require human data analysts. However, I do think there are some data analyst positions that could be at risk. Data professionals who mainly spend their time pulling data and creating reports may consider trying to add value beyond the simple, repetitive work. While data pulling and report building are important tasks, they are easy to automate, and it would be wise to expand your scope of work. #dataanalyst #dataanalytics #dataanalysis #ai #artificialintelligence #generativeai #chatgpt #data

AI is Going To Kill Data Engineering Jobs? (My Learnings 👇🏻) When I started my career in Data Science/Machine Learning at that time AI was mainly focused on specific tasks like predicting outcomes, categorizing images or text, and forecasting future events. These tasks often used simpler methods, like linear regression or decision trees, which many people learned about when they first got into data science. But then, large language models (LLMs) like ChatGPT came onto the scene 🤖 These models are a lot more complex and versatile. Instead of just one type of output, they can write text, answer questions, generate ideas, and much more, all while understanding and generating human-like text. The rise of these tools has led to many new startups and developments in AI, with some claiming these technologies might replace certain jobs, including in software engineering. While it's true they're powerful, they're also tools to assist and enhance our work, not necessarily replace us. They help automate repetitive parts of jobs or tackle complex tasks more efficiently, potentially freeing us up for more creative and strategic roles. I started using tools like ChatGPT for my day-to-day work and here are my learnings: https://lnkd.in/gXkrDGHs Are you using AI tools for your work? What are your thoughts on it? 👇🏻 #dataengineering #dataengineer #ai #llm

Dear friends, This week, I'd to share some thoughts in response to a question I’m increasingly receiving from manufacturing clients: "𝗛𝗼𝘄 𝗰𝗮𝗻 𝗺𝗮𝗻𝘂𝗳𝗮𝗰𝘁𝘂𝗿𝗲𝗿𝘀 𝗲𝗳𝗳𝗲𝗰𝘁𝗶𝘃𝗲𝗹𝘆 𝘀𝗰𝗮𝗹𝗲 𝗔𝗜-𝗱𝗿𝗶𝘃𝗲𝗻 𝗽𝗿𝗶𝗰𝗶𝗻𝗴 𝗮𝗻𝗱 𝘀𝗮𝗹𝗲𝘀 𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗿𝗶𝗴𝗵𝘁 𝘁𝗮𝗹𝗲𝗻𝘁?" After successfully implementing pilots for long tail repricing and sales next best actions using AI, many manufacturers find it challenging to scale these solutions into core competencies once the initial data science team departs. Here's a practical roadmap to achieve this transition: 🔧 1. Build an Initial Team: Data & AI Manager: Oversees processes and strategy. Data Engineer: Manages data pipelines and infrastructure. Data Scientist: Develops and refines algorithms. 🎯 2. Recruit the Right Talents: Within the Organization: Identify problem solvers who integrate coding into their tasks or create workflows with tools like Alteryx. Outside the Organization: Talent in manufacturing may be limited. Look for expertise in fields such as actuarial sciences, energy forecasting, or banking derivatives. 🔁 3. Establish Effective Routines: Generate Python Code: Leverage tools like ChatGPT but ensure you have robust quality checks and technical proficiency to enhance prompts. Manage Data Securely: Use secure platforms for data management. Create Workflows and Quality Loops: Establish routines to maintain and improve processes. 📊 4. Seamlessly Integrate AI-Based Optimization: Ensure that AI-driven pricing and sales optimizations are fully integrated into your end-to-end processes to maximize efficiency and effectiveness. By following these steps, manufacturers can transform AI-driven price and sales optimization into a sustainable core competency. I invite you to share your thoughts and experiences to enrich this discussion. #AI #DataScience #Manufacturing #TalentManagement #DynamicPricing

My Full-Time Job is Figuring Out How to Make Money For Myself With Tech: Here’s The One Big Lesson Every Beginner Needs to Know to Navigate the Future of Technology and Make Smart Decisions With AI First a confession: Admitting that AI can be truly helpfully creative was a challenging concept for me to accept. Yet, AI's creative ability is undeniable, as demonstrated by the sublime awe-inspiring text you are reading now (lol, I’m writing this. I promise…but AI helped). The Old Days (aka Yesterday): The other day, I was supposed to enter some set-up instructions into the shell terminal of my cherished sleek space-gray MacBook Pro, fondly named The Hammer (because with it, every problem is, indeed, a nail) but I did something different. Today With AI: Rather than blindly copy/paste the magic incantations into my mac, I blindly pasted the instructions into ChatGPT. No fancy prompting. Just plain-old-vanilla copy/paste. And in that one move, I learned more about my Mac’s cryptic black terminal window than I have in years of typing “<command> --help” or “man <command>”, Googling or searching Stack Overflow. The Future With AI (aka Next Few Months): AI chatbots don’t just make coding faster; they increasingly make coding accessible to non-techie (e.g., business, or creative) people who wouldn’t have dreamed about deciphering cryptic tech tools. On the other hand, I’m beginning to use AI tools for business and creative media stuff that's getting pretty darn interesting. The One Big Lesson: I will probably never replace any world-class analyst/manager, marketer or creative…and they may never replace any world-class tech professionals, like me. But it's not a zero-sum game. We will all just create better stuff for ourselves and our customers, friends, families, employers. I'm documenting my journey on using emerging technology to create a job for myself earning 340k, my starting offer as a Principal Solution Architect at AWS. Follow me to learn more Ola Adedoyin

🚀 The Rise of Generative AI: From 2018 to 2022 & The Power of Prompting! 🤖 The evolution of Generative AI has been phenomenal, and in today's post, I’m sharing an insightful flowchart showcasing ChatGPT's growth from 2018 to 2022. Dive into the process of prompt engineering and how the combination of prompts, LLMs (Large Language Models), and outputs is revolutionizing the field of AI. If you're a Data Analyst, Data Scientist, Data Engineer, ML Engineer, or AI Engineer, mastering the art of prompting will give you an edge in harnessing AI's full potential! 📈 Unlock the power of AI and take your skills to the next level! 💡 #rabdatamagic #GenerativeAI #ChatGPT #PromptEngineering #AIRevolution #MLEngineer #DataScience #AIEngineer #LargeLanguageModels #TechGrowth #AIinBusiness #DataEngineering #MachineLearning #ArtificialIntelligence #DataAnalytics #FutureOfAI

The single biggest reason why I believe AI will change organizations forever - is that everyone in a company and their grandmother / kids can relate to it. Cloud, Data warehouses/lakes, Analytics, data science, Machine learning - they are all technologies where the average CEO or middle manager cannot even approach to understand or “feel” the mindblowing-ness (not just use). And techies and tech companies mostly sound like “trust me, bro!”.. just hire techies, pay them 3x your average salary, build a new culture and look cool. 3 easy steps. AI, on the other hand, was made approachable. Part of every day lingo (even though ChatGPT is clunky to say every time). Use it for the most boring things. But works so freaking well. The average middle manager has felt ChatGPT (even if it’s to help her kids homework), experiences it firsthand that she can actually be productive. Without waiting for a techie or be confined to changing her behavior to use tech (“transformation”, anyone?) Now the conversation naturally goes to how do I do this org wide? Then leaders talk about “hey my data / infra is not ready.. but I see value. How can I do this?” That to me, is probably the single most important things that LLMs and GPTs have done. They flipped the conversation. #AI #LLMs #ChatGPT #transformation #digital #artificialintelligence