Mapping uncertainty using differentiable programming Uncertainty quantification (UQ) and propagation is a ubiquitous challenge in science, permeating our field in a general fashion, and its importance cannot be overstated. Recently, the commoditization...

We just released a new differentiable‐programming framework for rapid uncertainty quantification—no Monte Carlo required! 🚀 Check it out: doi.org/10.1002/aic.... #ProcessEngineering #UQ #DifferentiableProgramming

I see room for improvement. Proposal 🚀

5️⃣ Have more questions? Drop them anytime.
Chemical engineering might be more you than you think. 💬👋

#CollegeBound #ParentingTeens #AskAnEngineer #FutureReadyKids #STEMForParents

3️⃣ “Will I still have time for a life?”
Yes! ChemE students join clubs, play sports, travel, and still make time to rest. 🎭🌍😴

4️⃣ “Can I still go to med/law/business school?”
Definitely. ChemE is a rock-solid launchpad to almost any future path. 📊⚖️🩺

1️⃣ “Do I have to be a genius to study ChemE?”
Nope. Just curious, hardworking, and ready to learn new things. 🧠✨

2️⃣ “Is it just chemistry?”
Not at all! It includes physics, math, biology, computer science — and creative thinking. 🧪💻📐

4️⃣ Practice asking why and how. That kind of thinking is exactly what ChemEs do every day. 💭🔬

5️⃣ Start now. Start small. But start!
The world needs your ideas. 🌎🚀

#HighSchoolSTEM #FutureEngineer #STEMEducation #STEMTeachers #ExploreEngineering

3️⃣ Look for:
💥 Summer programs
💥 Lab tours
💥 Virtual events
💥 Competitions or science camps

4️⃣ Practice asking why and how. That kind of thinking is exactly what ChemEs do every day. 💭🔬

1️⃣ You don’t have to wait until college to start thinking like a ChemE. 🧠
Here’s what you can do now. ⬇️

2️⃣ 🔹 Take chemistry, physics, and calculus if you can
🔹 Join STEM clubs or science fairs
🔹 Try DIY science projects
🔹 Follow science creators online

4️⃣ The best part? It’s like real-life problem-solving with a side of experimentation. 🧠⚗️

5️⃣ If you love a challenge and like to tinker — you’ll love chemical engineering. 🧩🔥

#STEMisFun #ScienceRocks #STEMspiration #HighSchoolScience #NextGenEngineers

2️⃣ Think:
✅ Color-changing reactions
✅ Giant lab equipment
✅ Building prototypes
✅ Solving mysteries with science 🔍🔬

3️⃣ You’ll break stuff, fix stuff, build cool things, and maybe even invent something world-changing. 🔧💥💡

1️⃣ Wait — chemical engineering can be fun?
Absolutely. We said it. 🎉🧪

5️⃣ Whoever you are, there’s a place for you in ChemE.
Bring your interests — and let’s build your future. 🧩🚀

#STEMForTeens #DreamBig #FutureSTEMLeaders #EngineerYourFuture #GirlsInSTEM

3️⃣ Chemical engineering fits a wide range of personalities.
Creative. Quiet. Analytical. Social. Driven. Curious. 🌈✨

4️⃣ It’s not about being one “type.”
It’s about using your strengths to make something amazing. 💪🔬

1️⃣ There’s no one kind of chemical engineer — and that’s the best part. 🧠🎨

2️⃣ Are you:
🧪 The Experimenter?
🌱 The Planet Saver?
💡 The Inventor?
📊 The Strategist?
🎮 The Tech Tinkerer?

The 2026 Chemistry Faculty Jobs List

This is for chemical engineering faculty positions. There is another one for Chemistry positions: docs.google.com/spreadsheets...

AY25-26 Chemical Engineering Faculty Jobs List

The 2025-2026 faculty job list is at docs.google.com/spreadsheets.... It is currently pretty sparse. If your department is hiring this year, consider adding it to this list!

4️⃣ It’s not just about your first job. It’s about options for your whole life. 🔓📊

5️⃣ Parents: ChemE is one of the most flexible and respected STEM degrees.
Students: You can do anything with this background. 💸🌟

#STEMCareers #CollegePlanning #FutureEngineer #ParentsOfTeens #CareerExploration

2️⃣ ChemE grads consistently rank among the top-paid engineers — and they work in dozens of industries. 📈💡

3️⃣ With a ChemE degree, you can go into:
✔️ Tech
✔️ Medicine
✔️ Business
✔️ Policy
✔️ Energy
✔️ R&D
✔️ Startups

1️⃣ Want a career that’s meaningful and financially solid?
Chemical engineering delivers. 💼💰🌎

Spin-informed universal graph neural networks for simulating magnetic ordering | PNAS The screening and discovery of magnetic materials are hindered by the computational cost of first-principles density-functional theory (DFT) calcul...

👏 Congratulations to our alumnus Wenbin and collaborators, whose paper "Spin-informed universal graph neural networks for simulating magnetic ordering" is now online on
@pnas.org
! 👏

🔗Paper link: www.pnas.org/doi/10.1073/...

📝Paper digest 👇

The evolving role of programming and LLMs in the development of self-driving laboratories Machine learning and automation are transforming scientific research, yet the implementation of self-driving laboratories (SDLs) remains costly and complex, and

📚 Lower the barrier to automated science—Claude-Light offers a hands-on platform for students and researchers to learn automation, design of experiments, and AI-driven optimization. Explore it here: doi.org/10.1063/5.02...

Solving an inverse problem with generative models Inverse problems, where we seek the values of inputs to a model that lead to a desired set of outputs, are considered a more challenging problem in science and engineering than forward problems where ...

🚀 Generative models aren’t just for art—they can help scientists and engineers “solve the puzzle” of inputs and outputs. See how in “Solving an inverse problem with generative models”: doi.org/10.1039/D5DD...

The evolving role of programming and LLMs in the development of self-driving laboratories Machine learning and automation are transforming scientific research, yet the implementation of self-driving laboratories (SDLs) remains costly and complex, and

🌐💻 Ever wish you could chat with your lab instruments? Learn how LLMs can select devices, write code, and even call functions to run real experiments—all explained in this paper: doi.org/10.1063/5.02...

Solving an inverse problem with generative models Inverse problems, where we seek the values of inputs to a model that lead to a desired set of outputs, are considered a more challenging problem in science and engineering than forward problems where ...

🤩 Why gen-AI for inverse problems? It’s fast, flexible, and gives built-in confidence estimates. Perfect for science, engineering—or any “what-if” scenario. Check it out: doi.org/10.1039/D5DD...

The evolving role of programming and LLMs in the development of self-driving laboratories Machine learning and automation are transforming scientific research, yet the implementation of self-driving laboratories (SDLs) remains costly and complex, and

🎨 Say “I want this light output” and get the exact LED settings—no guesswork! See how generative AI and smart algorithms solve inverse problems in the lab: doi.org/10.1063/5.02...

Solving an inverse problem with generative models Inverse problems, where we seek the values of inputs to a model that lead to a desired set of outputs, are considered a more challenging problem in science and engineering than forward problems where ...

📊 We compare four methods—classic math, backward models, Gaussian mixtures, and diffusion flows—to tackle inverse problems. Surprise: AI-driven generators stack up beautifully! Dive into the details: doi.org/10.1039/D5DD...

The evolving role of programming and LLMs in the development of self-driving laboratories Machine learning and automation are transforming scientific research, yet the implementation of self-driving laboratories (SDLs) remains costly and complex, and

🤖💡 From simple Python loops to advanced AI agents, this paper shows how programming and large language models can turbocharge scientific discovery in “self-driving” labs. Read now: doi.org/10.1063/5.02...

Solving an inverse problem with generative models Inverse problems, where we seek the values of inputs to a model that lead to a desired set of outputs, are considered a more challenging problem in science and engineering than forward problems where ...

🎨 Imagine saying “I want this light intensity” and instantly getting the RGB settings. That’s exactly what our generative approach can do—conditioning on outputs to create inputs! Learn more: doi.org/10.1039/D5DD...