𝗥 𝗜 𝗖 𝗛 𝗘 𝗬 💡 𝗠 𝗔 𝗟 𝗛 𝗢 𝗧 𝗥 𝗔 AI Innovator | Tech Solution Builder | EdTech Specialist | Equality Champion Ｆｕｔｕｒｅ．Ｐｏｓｓｉｂｌｅ．Ｎｏｗ． 🚀 Empowering businesses and individuals to thrive in the AI-driven future. Whether you're a startup seeking AI growth, a company aiming to up-skill your workforce, or an aspiring tech professional, I provide tailored solutions to drive your success. 🚀 FOR ➜ STARTUPS OPTIMISE operations and accelerate growth with custom AI integrations. FOR ➜ CORPORATIONS DEVELOP inclusive tech training programs to enhance workforce capabilities. FOR ➜ INDIVIDUALS NAVIGATE career transitions into tech with personalised guidance and resources. ▶ Pioneering the convergence of AI, EDUCATION, and INCLUSIVITY. ▶ Specialising in CS education for novices and tech training for career changers. ▶ Dedicated to democratising tech knowledge, fostering diversity, and pathways for social mobility. ＰＲＯＦＥＳＳＩＯＮＡＬ　ＶＥＮＴＵＲＥＳ • 𝗔𝗹𝗳𝗮 𝗖𝗼𝗱𝗲: AI Full Stack Bootcamps for career changers • 𝗔𝘃𝗲𝗻𝗶𝗿 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀: Crafting bespoke AI/ML solutions for businesses • 𝗕𝗙𝗙𝗟 𝗕𝗹𝗮𝗰𝗸 𝗙𝗼𝘂𝗻𝗱𝗲𝗿𝘀 𝗙𝘂𝗻𝗱 𝗟𝗼𝗻𝗱𝗼𝗻: Empowering underrepresented founders • 𝗖𝗦𝟭𝟬𝟭 𝗖𝗼𝗺𝗽𝘂𝘁𝗲𝗿 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 𝗖𝗼𝗼𝗽𝗲𝗿𝗮𝘁𝗶𝘃𝗲: Revolutionising CS education • 𝗚𝗼𝗼𝗱 𝗕𝗼𝘆 𝗠𝗲𝗱𝗶𝗮: Amplifying diverse tech narratives and views • 𝗛𝗲𝗹𝗶𝗼𝘀 𝗧𝗮𝗹𝗲𝗻𝘁 𝗔𝗰𝗾𝘂𝗶𝘀𝗶𝘁𝗶𝗼𝗻: Bridging talent gaps in tech with an EDI focus 𝗥𝗘𝗔𝗗𝗬 🫱🏻‍🫲🏽 𝗧𝗢 🫱🏻‍🫲🏽 𝗖𝗢𝗟𝗟𝗔𝗕𝗢𝗥𝗔𝗧𝗘? Open to partnerships, and impactful projects. Let's connect and shape the future of tech together. #AIInnovation #TechEducation #CareerChange #CSForAll #EqualityInTech #SocialMobility Positions: Founder at Alfa Code (2022 - Present), Seed Investor at MRRM Labs (2024 - Present), Member Board of Trustees at Black Founders Fund London (2024 - Present), Interim Chief Technology Officer at Computer Science Cooperative (2023 - Present), Co-Founder at Helios Talent Acquisition (2023 - Present), Partner at Avenir Technology Partners (ATP) (2022 - Present), Podcast Host and Producer at Good Boy Media (2022 - 2023) Skills: Low-Code Development Platforms, Rapid Application Development (RAD), Pricing Strategy, Seed Capital, SaaS Development, Competency Based Interviewing, Education Policy, Media Production, Interviewing, Podcasting, Product Launch, Documentaries, Early-Stage Startups, Venture Financing, Social Entrepreneurship, Grassroots Organizing, Agile Application Development, Rapid Prototyping, Web Applications, Software as a Service (SaaS), Off-shore Team Management, Technology Recruitment, 360 Recruitment, Executive Search, Technology Leadership, Agile Leadership, Educational Technology, Higher Education Policy, Training Delivery, Curriculum Development, Diversity, Equality, and Inclusion (DEI), Digital Strategy, Co-sourcing, Trusteeship, Fundraising, Corporate Social Responsibility, Operations Management, Media Relations, Government, Sales, Equality & Diversity, Change Management, Team Leadership, Cross-functional Team Leadership, Leadership Development, Educational Leadership, Organizational Leadership, Team Building, AI and ML, Startups Recent Posts: Physics Informed Neural Networks (PINNs) are one of the coolest techniques in the last 5 years of Machine Learning. Why? Because PINNs combines scientific knowledge with the power of neural networks. It turns out that this is a very powerful technique to solve a large number of impactful problems ranging from detecting black hole gravitational waves to predicting Covid-19 cases globally. Physics Informed Neural Networks come under the broader field of Scientific Machine Learning. Scientific Machine Learning is an amazing new field which is fun, and equally powerful. Here are 3 resources to start with Scientific ML: (a) Scientific ML Book: https://sciml.ai/ (b) Julia's ecosystem for learning SciML: https://sciml.ai/ (c) World's first Scientific ML Bootcamp: https://lnkd.in/gPwRaj89 The animation shows how adding physics does much better than just a neural network! Animation credits: Ben Moseley P.S: Want to do research projects using this technique? Join our Scientific ML Bootcamp: https://lnkd.in/gPwRaj89 Is the AI's reasoning true intelligence, or simply reciting solutions from training material? New research by ByteDance (TikTok's owner) has shown that AI "exhibits extremely severe recitation behavior," failing tests if "conditions are subtly shifted." All vendor AIs suffer a 50%-60% performance loss on elementary problems by changing just one or two words in the question. For example, change "traveling" to "floating" for a question about boat distance calculation. Researchers used a benchmark consisting of pairs of problems: 1) original problems that AI can solve, and 2) same problems with a key condition modified so there's a different solution paradigm but less verbal modification as possible. All AIs' "modified" scores, including vision models, are 50%-60% lower than the original problem score. The long "thinking" process did not help. Can current confused stochastic parrots evolve to AGI or Super Intelligence and replace all human PhDs and software developers, as promised by some? Source: "Recitation over Reasoning: How Cutting-Edge Language Models Can Fail on Elementary School-Level Reasoning Problems?" arXiv 2504.00509. Meanwhile, AIs can look more human than humans in chats: "Large Language Models Pass the Turing Test", arXiv 2503.23674: "GPT-4.5 was judged to be the human 73% of the time: more often than interrogators selected the real human participant." ♻️Repost; 👍Like; 🔔Follow me for more. [you can hire me; or sponsor posts (3M views/month)] Grow your LinkedIn: https://lnkd.in/dD5BkzQc Today, we’re releasing the Perplexity Windows app. People have been asking for a Windows app ever since we released our Mac app. Our goal is to be present anywhere the user wants. This also helps us be a command shortcut away for any new query you want, to fight the search bar stronghold Google has on Chrome. You can dictate a new query by voice too. Download here and start asking! https://lnkd.in/gaK3XnBw 🌟 I'm thrilled to announce my participation in Perplexity Business Fellowship! This opportunity will allow me to deepen my skills in the fast-evolving field of AI. For those who may not be familiar, Perplexity is an innovative AI-powered answer engine that has truly transformed my workflow. I utilise it for grounding agent searches in multi-agent agentic systems, enhancing the ability to connect various agents and facilitate more informed decision-making. By leveraging its capabilities, I can ensure that each agent has the most relevant and contextually appropriate information, leading to improved collaboration and efficiency. Not only does Perplexity provide real-time answers from the web, but it also integrates organisational files and proprietary data from sources like Crunchbase and FactSet. Its new Deep Research feature is a game-changer, delivering expert-level analysis in just minutes, which is invaluable in the complex landscape of multi-agent systems. 🚀 Stay tuned for more updates on my journey and insights from this exciting experience! #Perplexity #AI #BusinessFellowship #DeepResearch #MultiAgentSystems #Innovation #Technology #ProfessionalDevelopment The power of knowledge! The #1 secret for managing a team: The Skill / Will Matrix 👇 ——— Danny Meyer is an iconic New York restauranteur. He's known for Shake Shack, Union Square Cafe, Gramercy Tavern, and others. But he also came up with this simple (but brutally-effective) 2x2 matrix for managers. It's the best framework to decide WHO on your team you should spend the most time with. He graphs two things: Willingness + Ability. ——— So, where on the chart should you spend the most time as a leader? Many people would say the bottom two quadrants ↳ the folks who need motivation or need to move on. Danny disagrees. He spends the MOST time with the top two quadrants ↳ the high performers and those who are untrained but coachable. Why? Because they are the key to long-term success. ——— And guess what... Danny doesn't just keep this matrix in his own head. He put it on the mirror of his employees' locker room. He's 100% transparent in how he manages the team. Every time his team gets ready, they can decide where they are now, and where they want to be. 🙌 ——— P.S. Know a great manager? Repost this so they'll see it ♻️. 📌 Want your own PDF of this? 1). Simply follow me (Jade Bonacolta) 2). And sign up for my free newsletter at TheQuietRich.co You'll instantly get this PDF (+ 9 others) directly in your welcome email! Escape the winter and fly to your dream beach destination. With the Cirrus SR Series G7, you set the course no schedules, no limits. Another year, another leap in LLM app development! 🪄 Our new 2024 LangChain State of AI report uncovers key trends in how developers are building LLM apps. In 2024, we’ve seen a shift from retrieval workflows to multi-step, agentic applications. 🤖 Open-source models are also gaining traction as devs seek more customizable infrastructure. Here's where the AI ecosystem is headed, based on data from LangSmith.👇 Read the full report: https://lnkd.in/g5yd92QW 🚀 Here is where you should start learning about Agents. - white paper by Google! ---------------------------------------------------------- 1. What is an Agent? 🧠 - Autonomous systems that observe, reason, and act towards achieving goals. - Operate via a cognitive architecture (model, tools, orchestration layer). 2. Distinction: Agents vs. Models - Models: Static, limited to training data. - Agents: Dynamic, access tools and real-world data, support multi-turn reasoning. 3.Cognitive Architectures 🛠️ - Enable agents to plan and adapt. - Frameworks like ReAct, Chain-of-Thought, Tree-of-Thought optimize agent decision-making. 4. Tools for Agents - Extensions 🛠️ Bridge APIs and agents for seamless tool usage. E.g., booking flights with Google Flights API. - Functions ⚙️ Enable client-side execution of tasks (e.g., processing API data asynchronously). - Data Stores 📚 Provide up-to-date information using vector embeddings for retrieval-augmented generation (RAG). 4. Applications - Use Cases:  Travel planning, smart home controls, data analytics, etc. - Implementation:  Combine tools, cognitive reasoning, and structured workflows for custom solutions. 5. Performance Enhancements 🚀 1. Techniques to boost agent intelligence: - In-context Learning: Adapt in real-time. - Fine-tuning: Train for domain-specific tasks. - Retrieval-based Learning: Use external databases for contextual insights. 2. Practical Examples - LangChain Integration 📜 Demonstrated an agent leveraging APIs to fetch sports schedules and venue details. - Vertex AI Agents 🔧 Production-grade tools for managing, scaling, and refining agents. Follow Lighthouz AI (YC S24) to learn how AI agents shaping up procurement! Do you want to get smarter answers? 🙋‍♂️🙋‍♀️🙋