Engineering

Roadmap to Learn Agentic AI This roadmap breaks down the journey into 12 focused stages: – Grasp the core differences between traditional AI and autonomous agents – Build a solid foundation in ML, LLMs, and frameworks like LangGraph, CrewAI, and AutoGen – Understand how agents use memory, plan actions, and collaborate – Learn to implement retrieval-augmented generation (RAG) and adaptive reinforcement learning – Deploy agents in real-world scenarios with performance monitoring and continuous improvement If you're building AI that goes beyond chat interfaces, this roadmap will help you architect systems that are capable, contextual, and action-oriented. Feel free to save or share if you find it valuable.

The ability to create clarity when there’s no shortage of chaos, opinions, and competing priorities is a rare skill. In any reasonably competent company, this skill alone will help take you quite far, fairly quickly. Concretely, this means creating clarity on the main problems, clarity on the right solutions, and clarity on the action plan & priorities. Very few people can do this well even though most people possess the intelligence necessary to do it. This is because most people in the workplace have been conditioned to add more information, sound more clever, satisfy more stakeholders, and feign more precision & certainty than is possible. Few understand that clarity in a chaotic situation can only emerge from subtraction, never from addition. Clarity comes from communicating what stands out as most important, why it is most important, how it will be achieved, and last but not the least, giving people a way of thinking about why it is okay, even great, that we aren’t doing All The Other Things.

𝗗𝗼𝗲𝘀 𝗳𝘂𝘀𝗲 𝗽𝗿𝗼𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝘀𝗲𝗿𝘃𝗲 𝗶𝘁𝘀 𝗽𝘂𝗿𝗽𝗼𝘀𝗲 𝗶𝗻 𝗲𝗹𝗲𝗰𝘁𝗿𝗶𝗰𝗮𝗹 𝗰𝗶𝗿𝗰𝘂𝗶𝘁𝘀? No electrical system is perfect; more critical is the issue of disturbances or faults. Every circuit is designed to carry a specific amount of current, commonly called a full load amperage (or current) (FLA). Whenever we go over the normal operating current, it may lead to excessive heat. The generated heat is a means of fire outbreaks. Increasing current above the normal load is an overload, not necessarily a fault. In some instances, we should protect circuits against overloads and, as such, will interrupt the circuit. However, if the current increases more than 125% (typical) of the FLA, a preventive means is needed to control the circuit's operating condition. Using overcurrent protection devices (OCPDs) like a fuse or circuit breaker interrupts currents that exceed the full load current 𝗯𝗮𝘀𝗲𝗱 𝗼𝗻 𝗱𝗲𝘀𝗶𝗴𝗻 𝗽𝗿𝗼𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝘀𝗲𝘁𝘁𝗶𝗻𝗴𝘀. In this illustration, different fuse sizes are used to test current flowing through the same circuit, and you can see the response. Each fuse was able to interrupt or cut off continuous current flow. The circuit remained intact, and no fire was seen. Without a fuse (or protection of any kind) but a copper wire used in place of a protective device, there was excessive heat, which led to fire. Electrical circuits usually operate most of the time without issues since faults or disturbances are one-time events. As such, we may not realize the importance of protection until a fault occurs. Protect circuits at any cost and safeguard your health, safety, and expensive investment. #electricalcircuits #protection #fuse #experiment

Exciting updates on Project GR00T! We discover a systematic way to scale up robot data, tackling the most painful pain point in robotics. The idea is simple: human collects demonstration on a real robot, and we multiply that data 1000x or more in simulation. Let’s break it down: 1. We use Apple Vision Pro (yes!!) to give the human operator first person control of the humanoid. Vision Pro parses human hand pose and retargets the motion to the robot hand, all in real time. From the human’s point of view, they are immersed in another body like the Avatar. Teleoperation is slow and time-consuming, but we can afford to collect a small amount of data.  2. We use RoboCasa, a generative simulation framework, to multiply the demonstration data by varying the visual appearance and layout of the environment. In Jensen’s keynote video below, the humanoid is now placing the cup in hundreds of kitchens with a huge diversity of textures, furniture, and object placement. We only have 1 physical kitchen at the GEAR Lab in NVIDIA HQ, but we can conjure up infinite ones in simulation. 3. Finally, we apply MimicGen, a technique to multiply the above data even more by varying the *motion* of the robot. MimicGen generates vast number of new action trajectories based on the original human data, and filters out failed ones (e.g. those that drop the cup) to form a much larger dataset. To sum up, given 1 human trajectory with Vision Pro  -> RoboCasa produces N (varying visuals)  -> MimicGen further augments to NxM (varying motions). This is the way to trade compute for expensive human data by GPU-accelerated simulation. A while ago, I mentioned that teleoperation is fundamentally not scalable, because we are always limited by 24 hrs/robot/day in the world of atoms. Our new GR00T synthetic data pipeline breaks this barrier in the world of bits. Scaling has been so much fun for LLMs, and it's finally our turn to have fun in robotics! We are creating tools to enable everyone in the ecosystem to scale up with us: - RoboCasa: our generative simulation framework (Yuke Zhu). It's fully open-source! Here you go: http://robocasa.ai - MimicGen: our generative action framework (Ajay Mandlekar). The code is open-source for robot arms, but we will have another version for humanoid and 5-finger hands: https://lnkd.in/gsRArQXy - We are building a state-of-the-art Apple Vision Pro -> humanoid robot "Avatar" stack. Xiaolong Wang group’s open-source libraries laid the foundation: https://lnkd.in/gUYye7yt - Watch Jensen's keynote yesterday. He cannot hide his excitement about Project GR00T and robot foundation models! https://lnkd.in/g3hZteCG Finally, GEAR lab is hiring! We want the best roboticists in the world to join us on this moon-landing mission to solve physical AGI: https://lnkd.in/gTancpNK

Last week, I described four design patterns for AI agentic workflows that I believe will drive significant progress: Reflection, Tool use, Planning and Multi-agent collaboration. Instead of having an LLM generate its final output directly, an agentic workflow prompts the LLM multiple times, giving it opportunities to build step by step to higher-quality output. Here, I'd like to discuss Reflection. It's relatively quick to implement, and I've seen it lead to surprising performance gains. You may have had the experience of prompting ChatGPT/Claude/Gemini, receiving unsatisfactory output, delivering critical feedback to help the LLM improve its response, and then getting a better response. What if you automate the step of delivering critical feedback, so the model automatically criticizes its own output and improves its response? This is the crux of Reflection. Take the task of asking an LLM to write code. We can prompt it to generate the desired code directly to carry out some task X. Then, we can prompt it to reflect on its own output, perhaps as follows: Here’s code intended for task X: [previously generated code] Check the code carefully for correctness, style, and efficiency, and give constructive criticism for how to improve it. Sometimes this causes the LLM to spot problems and come up with constructive suggestions. Next, we can prompt the LLM with context including (i) the previously generated code and (ii) the constructive feedback, and ask it to use the feedback to rewrite the code. This can lead to a better response. Repeating the criticism/rewrite process might yield further improvements. This self-reflection process allows the LLM to spot gaps and improve its output on a variety of tasks including producing code, writing text, and answering questions. And we can go beyond self-reflection by giving the LLM tools that help evaluate its output; for example, running its code through a few unit tests to check whether it generates correct results on test cases or searching the web to double-check text output. Then it can reflect on any errors it found and come up with ideas for improvement. Further, we can implement Reflection using a multi-agent framework. I've found it convenient to create two agents, one prompted to generate good outputs and the other prompted to give constructive criticism of the first agent's output. The resulting discussion between the two agents leads to improved responses. Reflection is a relatively basic type of agentic workflow, but I've been delighted by how much it improved my applications’ results. If you’re interested in learning more about reflection, I recommend: - Self-Refine: Iterative Refinement with Self-Feedback, by Madaan et al. (2023) - Reflexion: Language Agents with Verbal Reinforcement Learning, by Shinn et al. (2023) - CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing, by Gou et al. (2024) [Original text: https://lnkd.in/g4bTuWtU ]

This is what progress looks like, and is another reason I'm optimistic about the energy transition. Scotland's first ever wind farm was re-powered last year, 30 years after it began operating. It now delivers five times more power – using half the number of turbines. Key details: ✅ Hagshaw Hill wind farm entered service in 1995 ✅ Originally 26 turbines (16 MW total) ✅ Re-powered with 14 new turbines (79 MW total) Not only that, but every blade from the original turbines was recycled into new construction material – replacing concrete, timber, and plastics. The local community fund will also receive a huge boost, rising to £400,000 per year – 26x its previous level. Wind turbine blade recycling has long been one of the trickier challenges for the industry, both technically and economically, but progress is accelerating quickly. RWE recently completed installation of the first recyclable blades at the Sofia offshore wind farm – the first time this has been done at scale in the UK. Clean energy technologies just keep improving, and we are now starting to address some of the trickier challenges like circularity. Circularity isn’t just about recycling waste – it’s about designing assets with their second life in mind.

The worst mistake employers make? Waiting for a resignation to offer a pay rise. By that point it's too late. The damage is already done. As uncomfortable as salary conversations can be (they shouldn't!). You need to advocate for yourself. Your employer won't give you a raise if you don't ask. Here's How to Have a Salary Conversations Like a Pro: 1️⃣ Set Clear Goals with Your Manager ↳ Define what success & progression looks like. ↳ Set KPI's that justify a pay rise later. 2️⃣ Have Regular Conversations About Growth ↳ Don’t wait for the annual review. Check in quarterly. ↳ Ask: “What can I do to be in the best position for a promotion?” Work on a plan together to upskill, get more responsibility & add more value. 3️⃣ Document Your Success ↳ Track wins, metrics & business impact. ↳ Use those numbers in your performance reviews. Instead of “I’ve worked hard” say: “I led [Project] which increased [Metric] by X% and saved Y hours.” 4️⃣ Promote Your Work (Without Bragging) ↳ Don’t assume people know what you've done. ↳ Present updates, share results, speak up in meetings. 5️⃣ Make the Ask (So It Feels Collaborative, Not Demanding) ↳ Timing matters. Make it an agreed time or in line with company reviews. Try: “Based on my contributions in [Project], I’d love to discuss salary progression. What would it take for me to reach [target salary]?” 6️⃣ Leverage the Market (If Necessary) ↳ If nothing is happening internally, go outside. ↳ Get an offer on the table to give you leverage. If your company won’t pay you what you deserve, another one will. Retention is cheaper than recruitment. ♻️ Repost to help people advocate for themselves. 👋🏼 Follow Dan Mian for more career insights.

"I like my job and my company, but my salary doesn’t feel right". Aisha had been working in her company for three years. She enjoyed her work. Her team liked her. Her manager was supportive. But every time she saw her salary, she felt unhappy. “I’m doing more work now, but my salary is still the same,” she thought. This happens to many people. They’re happy with their company, but not with their pay. Aisha decided to take it up. Here’s what she did (and what you can learn too): 1. She did her research. Aisha checked online to see what others in her role were earning. She made sure her salary request was fair. 2. She picked the right time. She didn’t just ask suddenly. She booked a proper meeting with her manager—at a time when things were calm at work. 3. She made a list of her work. She wrote down her achievements: A process she improved Clients she helped keep happy Extra tasks she had taken on This showed how she was helping the company grow. 4. She knew what to ask for. Aisha had a clear number in mind. Not too high, not too low—just right for her skills and work. 5. She practiced what to say. She talked through her points with a friend first, so she could speak clearly and with confidence. 6. She stayed calm and polite. During the meeting, she didn’t complain or compare. She simply explained her work and asked for a raise. 7. She talked about the future. Aisha also shared her plans to keep learning and doing even better work in the company. 8. She was ready to talk it out. Her manager didn’t agree right away. There was some back-and-forth. Aisha listened and stayed open to different options, like bonuses or new projects. 9. She followed up. After the meeting, she said thank you. This showed she respected her manager’s time. 📌 What happened next? A few weeks later, Aisha got a raise—and a new opportunity at work. 💡 What can we learn? If you like your job but feel underpaid, don’t stay silent. Make a plan, stay professional, and speak up—just like Aisha did. Hope you have liked the article on how to ask for Salary Increment. Follow Me Smriti Gupta for Career & Resume tips #salarynegotiation #career #leadership

Grid bottlenecks are a feature — not a bug — of the energy transition. For years, we viewed economics as the main hurdle to scaling clean energy. High costs for wind, solar, heat pumps, and storage dominated the conversation. But the world has changed. Thanks to extraordinary innovation and dramatic cost reductions in renewables and electrification technologies, the bottlenecks we face today are different. They’re no longer about whether clean energy is affordable — it is. Instead, the challenge is whether our energy systems can evolve quickly enough to integrate it. A recent Financial Times piece highlights this clearly: across Europe, the rapid build-out of renewable generation now outpaces the ability of grids to move electricity to where it’s needed. Curtailment, congestion, and long queues for grid connections already cost billions annually — and without decisive action, these costs will grow. This isn’t a sign of failure. It’s a sign of success. It means the transition is happening faster than the infrastructure built for the fossil era can handle. The rise of decentralised, variable renewables and electrified heating and transport requires a fundamentally different approach to planning — one that anticipates growth rather than reacts to it. The EU’s move toward more coordinated, top-down scenario building and cross-border grid planning recognises exactly this. Better alignment between countries and system operators, faster permitting, and prioritisation of critical projects are essential steps to unlock the full value of cheap clean energy. Because every euro lost to bottlenecks is not a cost of climate action — it’s a cost of not modernising our grids fast enough. The more successful we are in deploying renewables and electrification, the more urgently we must upgrade and expand our grids. Grid constraints are not a reason to slow down. They’re a reason to speed up the transformation of an energy system that was never designed for the technologies now powering our transition.