The Data Guy

I think we are done for today... All tests passed ✅️

So this is why they started acting up weird 😂 https://shega.co/news/commercial-bank-of-ethiopia-unveils-digital-loans-up-to-150000-birr

When did CBE start Digital Marketing😂 seriously...?

The problem we have with our Generation is that, we are afraid to be seen trying... most people don't even try out of ego... but they actually really broke... Thats not even logical eko! This is our time to Grind and risk more 💯

Devtopia is back. We talked about what we were up to for the last 6 months. We are planning to do one episode per two weeks. We also have an editor now, Bisrat, so hopefully this will help us maintain our release schedule. [Devtopia Interlude E04]

The thing about Financial Services in Ethiopia is that Consistency is key... Was about to pay using Santim pay at a Cafe and the Waiters just told me that it won't work... bec customers has experienced system errors and lags.... Now just bec of few bad exp people won't even try... 🤷‍♂ Anybody tried to pay using Santim Pay recently...what was your experience...

For Those who might not know it, 365 Data Science is one of the best Places to learn Data Science and Data Analysis in details! Use this advantage well!

𝕋𝕙𝕚𝕤 𝕨𝕖𝕖𝕜 𝕛𝕦𝕤𝕥 𝕜𝕖𝕖𝕡𝕤 𝕘𝕖𝕥𝕥𝕚𝕟𝕘 𝕓𝕖𝕥𝕥𝕖𝕣🔥 365 Data Science has unlocked its entire AI & Data learning platform 100% free for 15 days (Nov 6–21). I took Data Strategy, Data Visualization, and Web Scraping courses last year and they turned out to be some of the most practical ones I’ve ever done. I was able to apply those skills directly at work, strengthen my analyses, and genuinely level up my career. If you’ve been planning to dive deeper into data or AI, this is your week. Now everyone can access: • 115+ AI and Data Science courses • CPE-accredited certificates • Real-world projects and practical case studies • Trusted by 2M+ learners globally No credit card. No catch. Just learning. 👉 https://365datascience.com/free-weeks-2025/ Credit: Miklol

For Anyone interested in VLM check this blog it will help you understand the implications simply https://towardsdatascience.com/how-to-apply-vision-language-models-to-long-documents/

If you have not seen this, you gotta check it out, there are Short Courses in there that can teach you many things, Power BI Included and get the Certification for free ❕

Next step is to build the MCP to do the design for you, and you will already be doing a fully Done Power BI

For Anyone who is working on Microsoft Power BI and is tired of Doing DAX and data modeling, Check this out... Its an MCP that you can Connect to your Ai model locally or on Cloud (like Claude) and just text what you want to do! It will literally Connect to you Power BI, do analysis and execute the code live and even do debug if wrong!!! Check it here

የማታ ሲሳይ 🎉 $250 from Antropic for free! Its for testing a product for them And I have until Nov 18 to finish! Can you guys recommend me what to build with it??

Write down which model you are using for which purpose?

Stop calling everything an LLM ❗️ Here are 8 model types you need to know 1️⃣ LLMs (Large Language Models) What they do: Generate text, answer questions what they don't: See images, take actions, run on your phone 2️⃣ LAMs (Language Action Models) These actually DO things. Book flights. Execute workflows Your LLM writes emails. LAMs send them 3️⃣ VLMs (Vision Language Models) 'What's wrong with this circuit?" *uploads photo* Spots the burnt capacitor AND explains why 4️⃣ SLMs (Small Language Models) Running on vour phone. No cloud. No latency. Your smart TV suggesting movies? That's an SLM 5️⃣ MoE (Mixture of Experts 100B) Parameters but runs like 10B Only activates the experts it needs. Like consultants, but useful. 6️⃣ LCMs (Latent Consistency Models) Real-time image generation. On your phone No waiting. No cloud. Just instant results 7️⃣ MLMs (Masked Language Models) BERT's children. They understand context from al directions. Your search engine's semantic understanding? MLMs 8️⃣ SAM (Segment Anything Model) Pixel-perfect object detection. Click once. Get perfect cutouts. Makes Photoshop's magic wand look like a crayon. see: The expensive mistakes → Using GPT-4 for image tasks (use SAM) → Cloud LLMs for edge devices (use SLMs) Building task automation with LLMs (use LAMs) → Running massive models for simple tasks (use MoE) We're in 2025. Stop calling everything an LLM. Start using the right tool for the job❗️ @datawithsimon

This is Your Week 😎 DataCamp just made 600+ premium courses completely free for one week 🚀 From November 3 to November 9, you can access everything from beginner to advanced tracks in Data, Al, and Tech - completely free. No credit card, no catch, just learning. I've used several learning platforms over the years, but DataCamp stands out for teaching the practical, job-ready skills that companies actually hire for. If you're serious about building or advancing your career in data and Al, start with one of these four tracks: Data Analyst Track: https://www.datacamp.com/certification/data-analyst Data Scientist Track: https://www.datacamp.com/certification/data-scientist Data Engineering Track: https://www.datacamp.com/certification/data-engineer With over 600 free courses, there's something for everyone - whether you're getting started or already in the field. If you're serious about leveling up in Data or Al, don't sleep on this. Opportunities like this rarely come free. Credit: Miklol

This is the current state of the Art STT and TTS service we have for our language! Make sure to join the channel! https://t.me/hasab_AI

Hasab AI Documentation is Live! Developers, this one’s for you — our official Hasab AI documentation is now available! You can now explore our APIs, learn how to integrate voice transcription, text to speech(TTS), summarization, and intent detection features, and start building with Hasab AI. 👉 Check it out here: https://developer.hasab.ai We can’t wait to see what you’ll build with Hasab.

🚀 I Built a Checkers AI Agent That Learns Like a Brain! 🤖♟️ Hello everyone! I just built a Checkers AI Agent that learns to play using the NEAT (NeuroEvolution of Augmenting Topologies) algorithm! 🧠♟️ After a long time of learning and building, I finally created a Checkers AI Agent that mimics biology using the NEAT algorithm. This AI doesn’t just play; it learns through self-play and evolves its neural network structure over time, just like a brain. Inspired by AlphaGo, I wanted to explore how AI can mimic biological evolution — and watching it get smarter is honestly exciting. It’s still at a basic level, but seeing it grow and learn on its own makes it incredibly rewarding. You can play against the AI in your browser or train your own models. It is also open source. Try it online, check out the code, or contribute! 🔗 Play now: https://checkers-using-neat1.onrender.com/ 🔗 GitHub: https://github.com/benasphy/Checkers_Using_NEAT ⭐ Don’t forget to star, fork. For questions or contributions DM @benasphy

Anyone who want to learn more on ML and Neural Networks, check this channel out !