A Business advocate with through Technology & Automation experience. Privileged to passionately serve at a Multinational Forbes Fastest 500 Software firm, A Business Intelligence Product startup and Currently spearheading data driven culture at NBP Fund Management Ltd. Proactive problem solver with excellent communication abilities. I am ambitioned to utilize and create unique applications of advance analytics and automation to create data driven solutions that focus on quantifiable business results. Qualified ACCA with a Honorary Bachelors degree awarded from Oxford Brooke's University. My continuous drive to learn and implement cutting edge technologies have brought me to earn a host of certifications around Artificial Intelligence, Data Science and Business Intelligence including a Tableau Desktop Specialist title from Tableau. On a voluntary contribution I head Karachi's Largest AI community labelled as Karachi.AI with 4K+ Professionals, Students and Academic Researchers. Welcome to my profile, Let's make this connection fruitful! Positions: Vice President / Senior Manager Business Intelligence at NBP Fund Management Limited (2020 - Present), AVP - Business Intelligence and Analytics at NBP Fund Management Limited (2019 - 2020), Manager Business Intelligence at NBP Fund Management Limited (2018 - 2019), Founder at Karachi.AI - Community of AI Practitioners (2017 - Present), Community Builder at Amazon Web Services (AWS) (2021 - Present), User Group / Community Leader at Tableau Software (2019 - Present), Business Intelligence Analyst at Modemetric (2017 - 2018), Associate Business Intelligence Analyst at Primatics Financial (2015 - 2017), Account and IT executive at MDC (2014 - 2015), Business Analyst ERP at CIS (Pvt) Limited (2014 - 2014) Skills: Onsite Training - Certified Data Engineer by Karachi AI, Generative AI, Amazon Bedrock, Big Data Analytics, Machine Learning, Management Accounting, Data Analytics, Data Science, Data Engineering, Tableau, SQL, Python (Programming Language), Business Analytics, Business Intelligence, Financial Analysis, Business Analysis, Financial Reporting, Finance, Research, Requirements Analysis, Agile Testing, Agile Methodologies, Enterprise Resource Planning (ERP), Data Warehousing, Data Analysis, Analytics, Business Intelligence (BI), Test Cases, Scrum, Extract, Transform, Load (ETL), Oracle SQL Developer, Microsoft Office, Microsoft Excel, QlikView, Watson, MySQL, Microsoft SQL Server, Java, Git, Team Leadership, Teamwork, Team Management, Data Warehousing & Data Modelling, MicroStrategy, Test Automation Tools, Business Intelligence Tools Recent Posts: Browser are changing for good, from just an interface to an agent that can not only load a website for you but also browse through it, action through it until the target is met. Community Members! Didn't register yet? https://lnkd.in/daM9T_BU The best series where you can learn a tool in a span of just 1 hour! We have an interesting topic to present in Ramzan Skill Series around developing Applications (Prototypes) with Prompting. During the current year we have seen reasoning models come-up and also we have seen AI getting too better at coding. We are covering both of them this Ramzan 1. Effective Prompting for Reasoning Models such as o1 or deepseek 2. Building Applications with (Bolt, Lovable or Vercel v0) If you have experience with any of the topics, please connect with us and we will honor you a 1 hour slot to train Pakistan's Largest and premier applied AI Community Karachi AI - Community of AI Practitioners CFBR / Repost please. Registrations are now live! Many of my connections have requested, everything happens in Karachi Onsite. This is the time of year where 8 Online 1 Hour Sessions, Free of Cost, in Ramzan, on Weekends are conducted after Taraveeh time 9 PM - 11 PM. You can still carry on leisure and me time and also learn skills too. (Link in the Post Description). Please comment "CFBR" a repost will help someone. Another success story out of Karachi AI - Community of AI Practitioners Mr. Ahmed Hassaan Qadri who is Non - CS. Karachi enabled him to be capable and attempt Google Cloud Skills Boost under #aiseekho program in collaboration with GDG Kolachi. Thanks Huzaifa Habib! From today, Spotify is now accepting audiobooks narrated by ElevenLabs Voices. More people are listening to audiobooks than ever before. But for many authors, getting an audio version of their work published on major platforms has been expensive and time-consuming — until now. Authors using ElevenLabs can now publish directly to Spotify and other major retailers through our new partnership with Findaway by Spotify — making distribution faster, easier, and more accessible than ever. Read more in The Verge: https://lnkd.in/eWZAGBGt And on our blog: https://lnkd.in/emrRguiz As ambassador to this wonderful community, I encourage connections to join AWS Community Pakistan Meetup in Karachi on Saturday 22nd Feb 2025. Link to event : https://lnkd.in/dE2WFxqM #AWS #generativeai Comment AI Expert! or Repost / CFBR Comment to share with people who can contribute. Note this is Online, so you can tag any one from Pakistan. Ever spent hours untangling a messy JSON file, scrolling endlessly just to find that one key piece of data? 😩 In AI & ML, raw data is everywhere, but making sense of it is the real challenge. 90% of information sent to the brain is visual, and visuals are processed 60,000x faster than text. 🚀 That’s why data visualization is a game-changer! 🔍 Meet JSON Crack – an open-source tool that transforms complex JSON data into interactive, intuitive graphs in seconds. Unlike plain text-based viewers, it helps you see relationships, patterns, and structures effortlessly. 🔥 Why It Stands Out: ✅ Interactive Graphs – Instantly visualize & explore JSON data. ✅ Multi-Format Support – Works with JSON, YAML, XML, CSV & more. ✅ Seamless Integration – Available as a web app & VS Code extension. With 36K+ GitHub stars, JSON Crack is becoming a must-have for developers & data scientists. Faster debugging, better collaboration, and clearer insights—all in one tool! 💡 How do you visualize data in your AI projects? Tried JSON Crack? Drop your thoughts below! ⬇️ Aykut Saraç #DataVisualization #AI #MachineLearning #JSON #OpenSource #DataScience Large Language Diffusion Models introduce LLaDA, a diffusion model with an unprecedented 8B scale, trained entirely from scratch, rivaling LLaMA3 8B in performance. A text generation method different from the traditional left-to-right approach Prompt: Explain what artificial intelligence is Calling out all AI & Data Experts or someone who has practically leveraged AI for a task that can be taught in an hour. Comment below will dm you details. Please comment "CFBR" for outreach. It was a great experience to learn deeper about Google cloud Big Query ML! Thanks to collaboration of GDG Kolachi & Karachi AI - Community of AI Practitioners #aiseekho Its a game changer for Businesses who want to soft experiment Machine learig over their data without too much technical debt at POC Stage. The pipeline it self take cares of Transformation of categoricals, but still flexible to provide transform API via SQL. Most of the structured Data is in Databases and BigQuery allows you to efficiently process the data in ML pipeline within database without leaving for a silo of Private Jupyter Notebook. Here is a quick cheat sheet! Working with BigQuery ML? Here’s a quick reference guide to keep you on track! 🔹 Label: Define a column as label or specify in OPTIONS using input_label_cols. 🔹 Feature: Features are passed through your SQL SELECT statement. 👉 SELECT * FROM ML.FEATURE_INFO(MODEL mydataset.mymodel) 🔹 Model: An object created in BigQuery that lives in your dataset. 🔹 Model Types: Supports Linear Regression, Logistic Regression, and more! 👉 CREATE OR REPLACE MODEL <dataset>.<name> OPTIONS(model_type='<type>') AS <training dataset> 🔹 Track Training Progress 👉 SELECT * FROM ML.TRAINING_INFO(MODEL mydataset.mymodel) 🔹 Inspect Model Weights 👉 SELECT * FROM ML.WEIGHTS(MODEL mydataset.mymodel, (<query>)) 🔹 Evaluate Model Performance 👉 SELECT * FROM ML.EVALUATE(MODEL mydataset.mymodel) 🔹 Make Predictions 👉 SELECT * FROM ML.PREDICT(MODEL mydataset.mymodel, <query>) BigQuery ML makes it easy to run ML models using just SQL! Do you want to learn more? https://goo.gle/aiseekho Congratulations Wasiq Niaz to cross milestone. Huzaifa Habib fruits have started to bear. Dishonest charts are coming from all directions, and they're only going to hit harder. Unless we push back. So I made an interactive guide for Defense Against Dishonest Charts: https://lnkd.in/gmirPtg4 The interaction makes it obvious how small changes can shift your point of view, and how the dishonest use that to their advantage. Maybe a baseline starts a little too far up, or a trend gets cherrypicked to "prove" a point. Once you know the tricks, it's a lot easier to counter with honesty and reality. What a speech, remarkable literacy of @narendramodi. I envy our leadership has such detail into AI. Shaza Khawaja you can change the label. We have talent and passion! What will be the Impact of Deep Research on the Research Community? Will PHD be still relevant as it has been? Boiler Plate Researchers: Trait: Those who deal with copy-paste of online plethora Outcome? - Replaced by the Automation of Intelligence (OpenAI Deep Research) Context Aware Researchers: Trait: Those who understand the current context and use excellence of Human brain to drive contextual research and meaningful outcomes. Outcome? - They will consume Google Deep Research tool and enhance their productivity and knowledge consumption to improve their outcome. With Generative AI on the corner... You either Surf the wave or get ready to be drown away :) What is your opinion on this! Share in comments. Hit the share/repost too. We are bootstrapped and we've built an Urdu LLM which outperforms leading LLMs. That too for less than $100. For the longest time, low-resource languages have faced challenges with open-source LLMs. But not anymore! Introducing the world's best URDU LLM (in our opinion and a lot of benchmarking ). Launching other languages soon! Our model beat Meta's Llama 3.1, Google's Gemma 7B, Mistral AI's Nemo 12B and Cohere's Aya model across various benchmarks. (we are still testing) This is what bootstrapped and open source looks like. Come change the world with us! Model Card: https://lnkd.in/g2TEhEA8 Wonderful experience completing Google Vertex AI Cloud Lab in #aiseekho partnered with GDG Kolachi. Outreach partner Karachi AI - Community of AI Practitioners Want to try? Link in comments! this is amazing 🔥 Urban Design Meets AI: Real-Time Text-to-3D Prototype Released STF Labs has unveiled UrbanGPT, a prototype combining ChatGPT and Grasshopper for real-time, text-driven urban design. This tool allows designers to: - Generate 3D city layouts by simply typing commands. - Track essential metrics like floor areas and program on the go. - Create realistic massing distributions, rendered instantly with AI diffusion technology. Currently, UrbanGPT is powered by GPT-4o, though it's adaptable to other language models. This advancement could streamline urban planning and design workflows, making processes more efficient and responsive. What are your thoughts on the impact of AI-driven design in urban planning? Paper: https://lnkd.in/dDWRhJbr Project page: https://lnkd.in/d-scbDKY #UrbanDesign #ChatGPT #AIInnovation #ArchitectureTech Read if you're associated with E-Commerce in any capacity. Pakistan eCommerce Association Daraz Shopify Perplexity Things are moving fast! Welcome Muneeb Qureshi let's do this. BREAKING: OpenAI just demoed an AI Sales Agent in Tokyo. It’s built to automate tasks that sales assistants handle daily: - Processing inbound leads - Enriching data - Scheduling meetings There are 3.1M sales assistants in the U.S. (Earning around $45K per year) That’s 1% of the population.  A massive market for automation. The AI follows a structured workflow: 1. A lead fills out a 'Contact Sales' form. 2. The AI processes the details and enriches the data. 3. It checks calendar availability. 4. It drafts and sends meeting invitations. AI adoption in sales is already reshaping the industry: (a few recent statistics on that) - Teams using AI are 1.3x more likely to increase revenue. - 70% of a sales rep’s time goes to non-selling tasks—AI reduces that. - 93% of enablement tech users are investing in AI. What’s still unclear about this leak: Is this a product for API users?  Or a direct enterprise solution? If AI sales assistants become the norm, startups in this space will need to innovate fast. We’re watching the next phase of AI in sales. It’s going (super) fast lately. PS: I’m Noam, posting daily about sales/social selling. BREAKING! OpenAI just made search open to EVERYONE, no account needed! Ironically, hours earlier, Google dropped the (stunning) Gemini 2.0. OpenAI is becoming Google, Google is becoming OpenAI. Here's what to know: While Google isn't getting the breathless press, it's probably the bigger announcement here - Gemini 2.0 is awesome, check out the capabilities below! Okay, let's do this: THE TWEETS: Sam Altman's tweet just said Make Search Great Again which is both lol and also again shows the looming US administration's influence here. Hours earlier, Sundar Pichai of Google tweeted about Gemini. Very cool. THE FACTS: OPENAI: OpenAI has made search it freely available to everyone without requiring an account or login as of today. It's a search engine now. Full stop. This marks the final step in a gradual rollout that began with: October 2024: Initial release to paid subscribers December 2024: Extension to all logged-in free users February 2025: Complete removal of login requirement GOOGLE: > Gemini 2.0 Flash: Now generally available through Google AI Studio and Vertex AI, this is what they're calling the workhorse model - it features a 1-million token context window and is optimized for high-volume tasks. Image generation and text-to-speech capabilities are coming soon. > Gemini 2.0 Pro Experimental: Google's most powerful model yet, with a 2-million token context window and enhanced coding capabilities. It's available in Google AI Studio, Vertex AI, and for Gemini Advanced users. > Gemini 2.0 Flash-Lite: The new budget friendly model that outperforms the previous 1.5 Flash while maintaining the same speed and cost. It can process around 40,000 unique photo captions for less than a buck. > Gemini 2.0 Flash Thinking Experimental: A special version that breaks down complex prompts into steps and shows its reasoning process, like what we like about Deepseek and ChatGPT's reasoning models. It can now integrate with YouTube, Google Search, and Google Maps. That should be interesting... All models support multimodal input with text output, and Google has implemented new safety measures using reinforcement learning to improve response accuracy. Fun day! ++++++++++++++++++++++++++++++ Looking to actually get fluent in generative AI? Generative AI for Professionals is all about behavioral change - it works. Thousands and thousands of folks agree. Find it on the AI Mindset website! If you want it for your enterprise, shoot me a DM! MY MIND HAS JUST BEEN BOGGLED. I just tried the early release of OpenAI's new DeepResearch feature (rolling out later today to the $200/month Pro users). I've been working in the CRM software industry for 30+ years. It's not just that I've had court-side seats to the game, I've been on the court, doing my best to play the game. First in vertical CRM (my first startup) and now as co-founder of HubSpot. I've had some modest success and I feel like I have a pretty good handle on things in the industry. That's why OpenAI's new DeepResearch feature boggled my mind. I asked it create a detailed research report including competitive analysis, positioning, growth, product strategy and AI vision for the industry. It thought about it for 11 minutes before responding. ELEVEN MINUTES! That's like an eternity in A.I. world. A.I. years are longer than dog years. What it produced was an 11,000 word report. With data. And citations. And tables. And genuinely great insights -- including some I hadn't really thought of before. What has me excited is not just that it can produce this kind of output (though that's pretty cool). What has me excited is that we'll be able to use this kind of output as *input* to a subsequent step in an agentic workflow. Because the future is about agent composability. Being able to pull together pieces and put them together into a larger whole. The same way we build teams to work together to tackle higher order missions and goals. This has been what I've been dreaming about for years. It's finally starting to happen. We are seeing more and more of the future -- and it's happening quickly. DISCLOSURE: I'm a small investor in OpenAI, and also a big fan. I've been helping with Adobe Project Concept. I think, it's one of the best use cases for AI. It's pretty much what I use. A browser-based mood-boarding tool that provides a flexible infinite canvas for ideation. Sign up for private beta here: https://lnkd.in/eD9MextY Karachi AI hosted Karachi AI Meetup # 22 : E-Commerce & Digital Marketing at Innovista Indus , powered by eocean. We are thankful to Mr. Danish Khalil who presented the topic "How AI is helping in E-Commerce Product Marketing and Advertisements" Here is a quick summary of his Talk! Danish Khalil's presentation centers on the application of Artificial Intelligence in enhancing e-commerce, particularly in product presentation and marketing. He emphasizes how AI can optimize various aspects of the e-commerce experience, and that AI is a useful tool to improve efficiency, personalize customer experiences, and boost digital marketing efforts. Here's a breakdown of the key areas covered: • Introduction to AI in E-commerce: Mr. Khalil defines AI in e-commerce as the use of advanced algorithms, machine learning, and data analytics to improve operations, personalize customer interactions, and enhance digital marketing. • Product Description and Labeling: Mr. Khalil discusses how AI can be used for various aspects of product information, including: ◦ AI-generated descriptions and listings. AI can assist in creating compelling and informative product descriptions. ◦ Tagging and organization. AI helps in categorizing and tagging products for easier search and management. ◦ AI Photography: The use of AI to generate product photos, which can save time and money on traditional photography. AI tools can adjust shadows and lighting, and allow for the use of reference images to guide the process. ◦ Generative Video Ads: AI tools can be used to generate captivating video ads by allowing the user to select a product and then the AI generates a video that can be further customized with templates, images, music, and animations. AI tools can also create UGC (User Generated Content) style videos for social media. Want a replay? : https://lnkd.in/dXEY3wyK Don't want to miss next one? : https://lnkd.in/dz4W9gkr Stay tuned as we honor to recap the event with detailed summaries of each talk! 🚀 DeepSeek AI R1 x Perplexity: When AI Reasons Like a Wall Street Analyst 🚀 I just put DeepSeek R1 through its paces on Perplexity with a complex challenge - creating an investment thesis for upcoming US IPOs. What happened next was fascinating. Instead of getting a generic response, I watched as R1 methodically: - Identified potential 2025 IPO candidates - Researched industry dynamics - Analyzed market conditions - Built a comprehensive investment framework The level of reasoning is remarkable - it's like having a brilliant research assistant who shows their work and thinking process in real-time. For Students and Perplexity Users: - Access R1 via "Reasoning with R1" under Pro settings - Perfect for research papers and complex analysis - Hosted on secure US/EU servers - Integrated with Perplexity's powerful search Here's the best part: Students get FREE access to all advanced features at pplx.ai/students 🎓 #AI #Innovation #DeepSeek #PerplexityAI #Research #StudentSuccess #Tech When the whole world was reeling under the impact of DeepSeek, another AI model from China just dropped in. Alibaba’s Qwen2.5-Max, released on Chinese New Year, is a Mixture-of-Expert (MoE) model. It has been trained on 20 trillion tokens, fine-tuned with RLHF and SFT, and benchmarked against top proprietary and open-weight models. Chinese companies have now tasted blood, and they are not going to slow down. Thankyou Saqib Intizar for your contribution. We will recap our talk in upcoming days from Karachi AI - Community of AI Practitioners Page. One of the best times! Resuming a learning Marathon again. The agenda US started for Open Source is picked up by China well. They (US) created a supremacy / novelty around closed Source selling has been doomed by the Open sourcing of Deepseek. For a long time censorship on Performing Chinese models have been imposed with ignoring them off the leaderboards. I don't see DeepSeek ias innovation or ground breaking unless prooven but still a miraculous win for the ecosystem. Now there will 1000x Deepseeks and many talented scientist will pickup from there on. Tokens will be free soon and the next chapter of Intelligence will be monetized. Reasoning is now free, which is fundamental to believe in AI response. True Multi Modal is also one of the next big frontier about to be demonetized. It is already but a phenonmal doom is about due. #NoBlackBox #Reasoning #Rationale 𝐂𝐡𝐢𝐧𝐚'𝐬 𝐑𝐨𝐛𝐨𝐭𝐬 𝐂𝐞𝐥𝐞𝐛𝐫𝐚𝐭𝐞 𝐂𝐡𝐢𝐧𝐞𝐬𝐞 𝐍𝐞𝐰 𝐘𝐞𝐚𝐫 𝐢𝐧 𝐒𝐭𝐲𝐥𝐞! While Tesla's Optimus is still tripping in the lab, Unitree Robotics' H1 robots are dancing in perfect sync at the Spring Festival Gala—flawlessly coordinated in front of millions. Tesla's robot takes 3 seconds to lift an arm. Unitree’s H1 completes a full upper-body loop in 0.8 seconds. Tesla can’t sync upper and lower limbs, while Unitree's robots move like a symphony. Speaking of symphonies—20 Unitree H1 robots performed with just a 0.8-second margin of error. Meanwhile, Tesla can’t even get two on stage together. The mechanical revolution is here.   Like what you see? Follow Kason Yu for insights on technology and cybersecurity. Click the 🔔 to get a notification so you don't miss my new posts. #ai #robots #innovation Thankyou Danish Khalil for your contribution. We will recap our talk in upcoming days from Karachi AI - Community of AI Practitioners Page. Thankyou Abbas Rajani for your contribution. We will recap our talk in upcoming days from Karachi AI - Community of AI Practitioners Page. Dear Connections, Haven't spoken on DeepSeek AI for a while which has been making waves lately with their open-source reasoning model, DeepSeek-R1. Many are hailing it as a game-changer, but here is my take on a balanced note. First off, while it's great to see more competition in the AI reasoning space, we need to keep our expectations in check. DeepSeek-R1 uses reinforcement learning (RL) to boost its reasoning skills—a solid approach, but not exactly groundbreaking. Tech giants like OpenAI have been refining this with techniques like Reinforcement Learning from Human Feedback (RLHF) for a while now. Then there's the model distillation angle. DeepSeek has slimmed down larger models into more efficient versions, which is smart, but again, not new. Google's Gemini project, for instance, has been down this road before. There's also some buzz about DeepSeek's training methods. Word is, they might have leaned on OpenAI's proprietary models during their training process, which could raise some legal eyebrows. If true, this could lead to some serious discussions about intellectual property rights in AI development. On the financial side, DeepSeek claims they built their model with about $6 million. That's pretty lean compared to the hefty budgets usually needed to train large-scale AI models from scratch. But if they're standing on the shoulders of existing models and techniques, those savings make sense. Still, it does make you wonder about the originality of their work. In above context NVIDIA 's fall seems to be opportunity to buy for Stock Market as I see the recovery very quick with new innovations in Blackwell Chips and more to accelerate AI to go beyond reasoning. In the end, while DeepSeek's efforts are a step toward making AI more accessible, we shouldn't mistake cost-cutting for true innovation. Affordable models are great for getting more people involved, but let's not confuse that with breaking new ground. + This is a balancer against US Domination or any other national domination on AI race, the World is a global village, World should dominate, No Race, No specific country and this is what DeepSeek's hype is about, Democratization! What are your thoughts on this? Is DeepSeek pushing the envelope or just repackaging existing ideas? Thankyou Iman Javaid 🇵🇸 for your contribution. We will recap our talk in upcoming days from Karachi AI - Community of AI Practitioners Page. Thankyou Hasan M. for your contribution. We will recap our talk in upcoming days from Karachi AI - Community of AI Practitioners Page. Deepseek just launched the official Janus-pro app Thankyou so much Speakers, Sponsors, Friends, Alumni and Community members to fill this auditorium of 160+ Since I’m reading many ill-informed takes, the situation with DeepSeek. They are not a small engineer team but one of the leading frontier lab (+100 researchers full time). They are not a newcomer. They started in 2023 by retraining a llama, pretraining, developing their internal data pipeline, then slowly rising to the top. I’ve been following them closely since February 2024. They made a large amount of innovations, most of them reproduced, both on the architecture and the data side. This last point is maybe even more crucially as western labs hardly talk about it: created their own OCR model+VLM for parsing scientific literature etc. It’s all documented in their 16 (!) papers. They trained the DeepSeek V3 model for 5 millions *in GPU cost*. Their article does not make any inference about total costs that would obviously include salaries, data work, tests and ablations. I do think this is a credible figure. Databricks trained a 36B active parameters MoE for 3K GPUs/$10M one year ago, so without many innovation. We are well within that range. One of the DeepSeek's main achievement is winning the chinese AI inference price war: they cut down price hard and were still earning money while their competitors running at loss. It makes it all the more credible they have a very efficient training pipeline. My overall take: DeepSeek is for me where Anthropic and OpenAI were two years ago. Very talented team with +100 researchers, exclusively focused on scientific research/engineering, with mid-level GPU resources but nothing gigantic. They are in the ideal sweet spot for top AI research: consistent, very focused not that large organization with some industrial level capacity. In contrast, US labs may have grown too much, too fast. Follow the page, as this time you will be able to read through all the speaker talks as summaries and ofcourse you can listen to them on youtube as well. Opportunity for Statistics Instructor with Karachi AI - Community of AI Practitioners Language: Python Medium Onsite : Karachi Experience : Atleast 3 years of Applied Statistics in Industry (No Bookish Teaching) , Needs to have practical exposure. Topics : Fundamentals, Hypothesis Testing , Regression, A/B testing and more.. Comment below if you're interested and fit per above criteria mentioned or know someone, please tag. I am thrilled to share that I have completed a comprehensive data analytics certification course and am now a certified data analyst! 🎓✨ This intensive 16-week course covered various topics including Python programming, statistics, machine learning, SQL, and Tableau. I had the privilege of studying at an institute affiliated with FAST University in Karachi, and I was part of the 9th batch to graduate from this program. The course not only equipped me with practical skills but also fostered a deep understanding of data science concepts. I am incredibly grateful to all my instructors (Saarah Rasheed, Usman Abbas, Umair Hanif, Mesum Raza Hemani) who worked tirelessly to impart their knowledge and guide us throughout the program. With this certification, I am now ready to embark on my data science research journey. The course has provided me with a solid foundation, enabling me to apply my skills in real-world scenarios and contribute meaningfully to the field. If you'd like to learn more about my experience or have any questions about the course, please don't hesitate to reach out to me. I would be more than happy to share additional information. Thank you Karachi AI - Community of AI Practitioners for your support, and here's to exciting opportunities in the world of data science! 🚀💡 #ai #karachi #machinelearning #experience #dataanalytics #community #data #programming #sql #tableau #datascience #python The world is changing fast Visual Electric to Runway for eyewear concept. Great for internal proof of concept. So quick (30 seconds) that I don't see why I wouldn't do this with a lot of concepts. Client presentation is a different story. Also, really liking the UI with Visual Electric. Those folks are doing a good job. Makes the tool fun to use. Hector AIxCreative - On a mission to ensure designers have a seat at the table in the age of Ai via AIxCreative! Newsletter - One Weekly Ai Asset: https://lnkd.in/guc_nyhU Game changer! 🚀 Calling all tech enthusiasts! The first-ever AWS Community Day Karachi is happening soon, and we’re looking for passionate volunteers to join us in making history! ✨ Why volunteer? • Work alongside tech industry leaders 🧑‍💻 • Get hands-on experience in a professional environment • Earn a volunteer certification to boost your profile If you’re eager to contribute to the tech community and be part of something BIG, this is your chance! 📅 Limited spots available. Register now and let’s build something incredible together! https://lnkd.in/dZ_JHK2d #AWSCommunityDay #AWSUserGroupKarachi #AWSCommunityDayKarachi Microsoft has released a new GenAI model called Trellis🔥🔥 All these 3D objects have been generated from a single input image.. You can now generate 3D models and assets with Trellis which is structured 3D latents for scalable and versatile 3D generation. It takes an input image or multiple images and reconstructs a 3D model within 10-20 seconds. TRELLIS is a large 3D asset generation model with 1.2 billion parameters and is by far state-of-the-art! It takes in text or image prompts and generates high-quality 3D assets in various formats, such as Radiance Fields, 3D Gaussians, and meshes. The cornerstone of TRELLIS is a unified Structured LATent (SLAT) representation that allows decoding to different output formats and Rectified Flow Transformers tailored for SLAT as the powerful backbones. High Quality - It produces diverse 3D assets at high quality with intricate shape and texture details. Versatility - It takes text or image prompts and can generate various final 3D representations including but not limited to Radiance Fields, 3D Gaussians, and meshes, accommodating diverse downstream requirements. Flexible Editing - It allows for easy editings of generated 3D assets, such as generating variants of the same object or local editing of the 3D asset. The output is achieved by integrating a sparsely-populated 3D grid with dense multi-view visual features extracted from a powerful vision foundation model, comprehensively capturing both structural (geometry) and textural (appearance) information while maintaining flexibility during decoding. They employ rectified flow transformers tailored for SLAT as the 3D generation models and train models with up to 2 billion parameters on a large 3D asset dataset of 500K diverse objects. 👉 Check out moe examples, project page, paper and code here: https://lnkd.in/dh_X4AU8 👉 I also teach my system, how to get started with freelance and become an AI Engineer in my AI Career Program: https://lnkd.in/dRm7xC6K