PYMUG (Python Mauritius User-Group) | LinkedIn (original) (raw)
Services et conseil en informatique
Location Not Relevent, Location Not Relevent 560 abonnés
Python User-Group for Mauritius
À propos
User Group of Python for Mauritius
Secteur
Services et conseil en informatique
Taille de l’entreprise
2-10 employés
Siège social
Location Not Relevent, Location Not Relevent
Type
Établissement éducatif
Lieux
Employés chez PYMUG (Python Mauritius User-Group)
Nouvelles
- PYMUG (Python Mauritius User-Group) a republié ceci
This week,I'll be speaking at the PyMUG March meetup. There is one subject I've grown passionate about recently-building retrieval sytems(aka the technology that powers chatbots and AI assistants). We're all familiar with the simple RAG architecture, but anyone who's played around with it knows you hit a ceiling very fast if you rely solely on naive semantic search (our beloved cosine similarity search). So I'll be focusing on modern RAG architectures,ranging from agentic RAGs (LangGraph-based), GraphRAG (exploiting knowledge graphs), and LightRAG (a newer variant). These newer systems have been proven to ensure completeness of answers with good precision. Towards the end, we'll also have a discussion on whether we still need RAG now that we have MCP tools to connect models to external systems. If this is your kind of thing,RSVP on this link: https://lnkd.in/dZQ24AiS #AI #RAG #GenerativeAI #ChatBots #LLM #DataScience #KnowledgeRetrieval #MCP
Pymug March Meetup, Sunday 15th at Coder Faculty, Curepipe On the menu: 📖 Rag systems(types of rag)+rag vs mcp (Urousha Jeetun) 📖 ZVec: Inside Alibaba's Newest Drop (Abdur-Rahmaan Janhangeer) RSVP: https://lnkd.in/dZQ24AiS - Pymug March Meetup, Sunday 15th at Coder Faculty, Curepipe On the menu: 📖 Rag systems(types of rag)+rag vs mcp (Urousha Jeetun) 📖 ZVec: Inside Alibaba's Newest Drop (Abdur-Rahmaan Janhangeer) RSVP: https://lnkd.in/dZQ24AiS
-
560 abonnés
2 mois Modifié
Yesterday at CoderFaculty in Curepipe, we had our first monthly PyMUG meetup for 2026.Nirmal R. showcased a very practical use case for MCP. “From Chat to Action: Practical Local Agents with MCP“ showed how the exciting new technology could be used to build reliable and practical workflows using LLMs and tool calls to perform basic data analysis. Dataset discovery, type inference, statistical profiling, with guardrails built-in. Built with Python, GPT-OSS 20b, and FastMCP.Dominique Theodore presented “Test the Boring Stuff with PyTest”, which was about PyTest and its powerful features such as markers, fixtures and parametrization that make testing more practical and less intimidating. Thanks to CoderFaculty for hosting and everyone who attended! 🔥#Python #PyMUG #Mauritius #Testing #AI #MCP - PYMUG (Python Mauritius User-Group) a republié ceci
THANK YOU!! WOW! We hit our initial fundraising goal of $314,159.26 🌠 We're bursting with joy and hope- and we're not done yet! Thanks to you, we’re rolling out two new goals: 400k (stretch) 512k (super stretchy!) Can you help us hit our super stretchy goal? Read on to learn how ⬇️ -- Grab PyCharm from JetBrains at 30% off before Dec 12 🐍💸 ALL proceeds go to the PSF! Already have PyCharm? Use discount code PYCHARM4PYTHON to renew your subscription at 30% off- ALL proceeds go to the PSF! Get PyCharm: https://lnkd.in/gMQH7D8f-- Become a PSF Supporting Member! Your membership helps keep Python strong, open, and for everyone- and gives you a voice in the future of Python and the PSF. Become a member today 🪪👉 https://donate.python.org/-- Make a donation! Your gifts help to: - Keep Python thriving - Support CPython and PyPI progress - Increase security across the Python ecosystem - Bring the global Python community together - Make our community more diverse and robust every Donate today 💝 https://donate.python.org/-- Donate to one of our incredible Fiscal Sponsorees! Your support helps keep these groups and #Python open and thriving for everyone. Donate today 👉 https://linktr.ee/thepsf-- Share your Python story! If Python has impacted your life, join our year-end fundraiser by sharing your story & linking to our donation page: https://donate.python.org/Every story matters—whether it simply brightens someone’s day or inspires a donation. Your story strengthens our community 💖 -- To wrap up: again, thank YOU. Our heartfelt gratitude to everyone supporting this campaign- even your re-posts help! It means so much to all of us at the PSF. We're incredibly grateful to be in community with you and to stand together behind this message: Python is for everyone.#PythonForEveryone - PYMUG (Python Mauritius User-Group) a republié ceci
Today with PYMUG (Python Mauritius User-Group) and DTG Labs I had the opportunity to deliver a workshop on the Basics of Gen AI and Applications of Gen AI in the industry to the students at University of Mauritius and Middlesex University Mauritius to prepare them for the upcoming hackathon next week. The topics basically covered were: 1. What is Gen AI? 2. How do LLMs like Chatgpt understands the meaning of texts. 3. What are Vector databases? 4. Retrieval Augmented Generation (RAG) - Application of Gen AI. It was a real success with the students being proactive during the session by asking the right questions and being motivated with passion to do so. One thing I was happy to see today: - Mauritius got talent! A big thank you to the mentors of this workshop: Thierry Lincoln Abdur-Rahmaan Janhangeer Dominique Theodore Tasweem Beelunkhan Ali Hoolash - who made this workshop possible by also delivering amazing sessions on AI Agents, Data and Machine Learning in the industry and the python frameworks and concepts. Also, much thanks to those who attended and participated (I could find only 10/23): - Akash Beedaysee- Ameera N. Manjoo - Ridwan Ghamy- Deepika Johaheer- Gaurav Hurrydoss- Harshita Awotar- Manvi Khushi Dabydoyal- Mitilesh Ramanna - Muhammad Talha Moossajee- Darshanaa Devi BeedassyFor those who are interested to learn more about Gen AI and its applications in the industry, feel free to contact Developers.Institute.Mauritius #AI #ArtificialIntelligence #GenAI #MachineLearning #Data #AIAgents #AIforBusiness #RAG
`` - PYMUG (Python Mauritius User-Group) a republié ceci
Today with PYMUG (Python Mauritius User-Group) I did a workshop about #AI #Agents like the one we are building at ai.leadscloser.com to University of Mauritius and Middlesex University Mauritius students to prepare them to the Upcoming Hackathon next weekend Other talks by the PYMUG teams included - Python Basics 1 & 2 by respectively Abdur-Rahmaan Janhangeer Dominique Theodore- Data Science & ML by Tasweem Beelunkhan- Embeddings, RAG, vector databases, LlamaIndex by Rayhaan DustagheerSpecial thanks for those who attended . All the best for the Hackathon . With these new tools, I made sure those students understand that if they have the builder mentality, they can achieve a lot and propel Mauritius tech ecosystem much further . Dr. Avinash Ramtohul, FBCS CMgrProud participants ( i could only find 10/16) : Akash BeedayseeAmeera N. ManjooDeepika Johaheer Dominique Theodore Gaurav Hurrydoss Harshita Awotar Manvi Khushi DabydoyalMitilesh RamannaMuhammad Talha Moossajee..Ridwan GhamyFor those attending , share what you though of it ?
`` - Interesting page that shares AI concepts for both tech and non-tech people.#GenAI #ArtificialIntelligence #Python #AI #MachineLearning https://lnkd.in/dJ_TcjeN
By 2026, AI fluency will be basic professional literacy. Every business decision, campaign, and workflow now carries an invisible layer of AI logic underneath. And you have the potential to lead the conversations that matter most in your industry. Learning how AI thinks is often more useful than knowing how it’s built. Knowing the basics is enough to join important conversations: • A sales lead uses RAG to generate accurate proposals from product docs • A lawyer uses prompt engineering to extract clauses from long contracts • A marketer fine-tunes an LLM to lock consistent brand voice • A founder uses embeddings to build a searchable company knowledge base • A teacher uses chain of thought prompts to improve student reasoning You can either sit on the side lines, or you can guide how your industry uses it. If AI handled your daily tasks tomorrow, would your unique contribution still stand out? Follow for more AI contents.#AI #GenAI #LLM #DTGLabs #AIAgents #MachineLearning #RAG #ArtificialIntelligence #GenerativeAI - Some good experience based insights on AI Agents:https://lnkd.in/d8Ufcjzk
101 abonnés
5 mois Modifié
A Year with AI Agents: What Worked, What Didn’t, and What’s Next At DTG Labs, we’ve spent the past year immersed in the world of AI agents experimenting, iterating, and scaling them from prototypes to production. Through building and deploying agents across real products and startups, we’ve discovered what truly works, what doesn’t, and where the technology still needs to grow. These are some of our biggest takeaways, not definitive answers, but lessons learned through experience that might inspire new ideas. What Is an AI Agent? An AI agent is a system that autonomously pursues goals and completes tasks on behalf of a user or organization. It doesn’t just respond to prompts, it observes, reasons, acts (via tools or workflows), and learns from results to continuously improve. Our Key Lessons 1. Frameworks Are Just the Plumbing We’ve tested CrewAI, LlamaIndex, LangGraph, and n8n. Frameworks are helpful, but what truly matters is your pipeline design: how data flows, how reasoning unfolds, and how tools connect. 2. Context Is Everything You can’t just give an LLM a goal and expect magic. The agent’s success depends on context: prompts, tools, memory, and environment. Well-structured context often beats sheer model size. 3. Simplicity Outperforms Complexity Our best agents were the simplest: clear objectives, focused prompts, and one or two tools. Over-engineering creates fragility. The best agents do one thing exceptionally well. 4. People Matter More Than Technology After working with multiple startups, one thing is clear: the human factor outweighs the technical. Culture, iteration speed, and clarity of vision determine success more than any framework or model. Conclusion We’re still early in the agent revolution. Soon, agents will quietly power nearly every product; automating workflows, personalizing experiences, and managing complexity in the background. At DTG Labs, we’ve seen both wins and failures, made mistakes, and learned valuable lessons. One truth stands out: Agents aren’t the product — they’re the enabler. If you’d like to explore how agents can transform your products or workflows, we’d love to collaborate. Let’s build the future together.#AI #AIAgents #AIinBusiness #DigitalTransformation #FutureOfWork #AIInnovation #AIAutomation #LLMs #DTGLabs #AIEngineering #MachineLearning #GenerativeAI - Good read for business introducing AI into their workflows.https://lnkd.in/djfesDkY
101 abonnés
5 mois Modifié
The Top 5 AI Challenges and what we Learned tackling them at DTG Labs AI has incredible potential, but building real world AI solutions isn’t just about models and code. It’s about solving deep organizational, ethical, and data challenges that decide whether your AI project succeeds or fails. After tackling multiple AI projects across industries, here’s what we’ve learned at DTG Labs. 1. Data Challenges AI is only as good as the data it learns from. We’ve seen: - Messy or inconsistent datasets - Data scattered across silos - Privacy and compliance risks Our approach: Clean, validated pipelines. Unified data layers. Strong governance and encryption from day one. 2. Ethical AI AI must work for people, not replace them. We bake Ethics-by-Design into every workflow, keeping a human-in-the-loop for critical decisions. Accountability, fairness, and transparency aren’t optional. They’re the foundation of trust. 3. Regulation & Compliance AI laws are evolving fast: from GDPR to the EU AI Act. We integrate compliance into the building process: - Audit trails - Documentation - Clear IP and data-usage rights Compliance isn’t a checkbox; it’s part of responsible innovation. 4. Bias in AI Bias hides in data and models and it can break trust instantly. At DTG Labs, we actively detect and mitigate bias through fairness audits, diverse datasets, and continuous monitoring to ensure every prediction is equitable. 5. Transparency & Explainability “Black box” AI doesn’t inspire confidence. We use Explainable AI (XAI) to make our systems understandable and auditable. When clients see the why behind a model’s decision, trust follows naturally. Our Core Belief Every AI challenge hides an opportunity: - Data issues → cleaner pipelines - Ethical questions → stronger governance - Bias → fairer systems - Regulation → smarter compliance TLDR: AI success comes from anticipation, not reaction. That’s how we build at DTG Labs where innovation meets responsibility. Let’s talk about how your organization can adopt AI responsibly and effectively. Message us to explore how we can turn your data into decisions — and your ideas into impact.#AI #DataScience #ResponsibleAI #EthicalAI #DTGLabs #ArtificialIntelligence #Innovation #AIConsulting #AIAutomation #AITransformation