Takeaways over information overload. I teach because I do not want my students to go through what I went through. Through most of my schooling in India, my education was predominantly lecture-based, or what one would call didactic. The teacher delivered, the student received, and most of the real learning happened later, alone with a textbook. When I came to the United States for my PhD, I took several courses where students were expected to think out loud and be wrong in front of each other. The difference in what I vs. classmates retain, was significant. The principle I have built on since is that takeaways matter more than information density. By mid-semester, undergraduates carry cognitive debt across five or six courses. Most take a class because it is required, only secondarily because the topic interests them. So rather than maximizing information per session, my job is to plant a small number of seeds that last more than a semester: ideas, habits, and ways of asking questions that will still be there when the equations have faded.

Three habits I want students to leave with. My field sits at an intersection of different sciences. Aerosols, for instance, are tiny-scale particles, yet they aggregate into phenomena that operate at scales thousands of kilometers wider. "X causes Y" is almost always a simplification of "X interacts with Y, which depends on Z, which is itself responding to factors we may not have accounted for". I want students to leave with at least the beginning of three things: 1) the habit of tracing those pathways, 2) the ability to estimate whether a process matters before reaching for its functional form, and 3) the understanding that every result is a value paired with an uncertainty, that deserves the same attention as the value itself. Uncertainties are part of every phenomenon, but the temptation to ignore them is often strong. This was quite apparent when I was teaching climate effects of aerosols, where most of the significant effects come with large uncertainty attached. They are an honest accounting of what we do not yet know and where the work of closing that gap should go next.

Fact-first inquiry. The method I rely on most is to state a result plainly and then question students back into it. In one session, I told the class that turning off the condensation process in a standard aerosol simulation increases the total particle number by a factor of 500. Then I spent the next ten minutes asking them why. What does condensation normally do to small particles, and what disappears when we turn it off? The fact acts as the anchor while the questioning builds the conceptual scaffolding underneath. One student wrote on the end-of-unit survey that the class "never felt like you were rushing through topics without checking our understanding first," which is precisely the type of learning I am after. What I look for in those ten minutes is a specific shift, such as a voice that was quiet becoming bolder, a student who was dazed now looking up and arguing back. I treat that shift, more than having a correct answer, as the real signal that learning is happening, which is why I consistently ask students to commit to a position out loud. It is in that shift that they genuinely internalize the concept rather than me just feeding them facts.

Inductive sequencing. Rather than leading with equations, what works better for me is to guide students inductively from physical observations to the formal theory. For instance, I ask why the sky over Tucson turns orange during wildfire season but turbid during high-humidity events. Highlighting this stark contrast in optical behavior motivates the underlying physics. Instead of passively receiving the formal theory, students first recognize a gap in their own understanding, making them eager for the theoretical framework that resolves it. This approach maps onto what is called "inventing to prepare for learning" or “creating a time for telling”, where students who grapple with a problem's structure before formal instruction, learn it more deeply.

Surfacing prior knowledge and misconceptions. I also want to know what assumptions students are already carrying. The CIRTL training emphasized the role of prior knowledge and misconceptions, which has been useful for me. A Slido word cloud gives me what students are thinking in real time and gives them a low-stakes way to commit to a position before discussion. One activity I used to open a lecture on modeling asked, "What specific information or ingredients do we need to model how long it takes to boil an egg?". The responses sorted almost perfectly into the three categories I wanted to teach: initial conditions (egg size, water salinity), boundary conditions (ambient pressure, pot size, heat source), and parameterizations (convection, heat transfer, composition of egg and shell). "Ambient pressure" came up largest on the word cloud, which then led to follow-up questions about why students had chosen those inputs. I pointed at their own answers and asked which they chose deliberately and which they assumed without saying so. By the time we reach the meat of the topic, the distinction between a specific input and an implicit assumption is no longer vague.

Bloomifying assignments. I extend the same design down to the question level. The CIRTL materials on Bloom's taxonomy pushed me to stop writing diffuse 'explain' prompts and instead target a specific level of thinking with a bounded answer space, such as a fill-in-the-blank, a binary choice followed by a one or two sentence justification, or a specific numerical calculation. For higher-order questions, I followed the strategy of building Analyze and Evaluate items around realistic data, asking students to interpret a size distribution or compare two simulation outputs rather than to recite a definition. The payoff in this approach is twofold. Students recognize what kind of thinking each question is asking for, leading to cleaner, more confident answers, and grading becomes faster and more consistent because each question now has a narrow space of correct responses.

Assessment philosophy. My instinct in assessment is more formative, with frequent, low-stakes checks that diagnose where students are in real time, rather than a few high-stakes graded events. The course I co-taught was already structured around a midterm, a final exam, a term project, and short in-class quizzes. Within that inherited structure, I redesigned the final term project from a literature review into a real-world data analysis exercise. That change pushed the summative model closer to authentic inquiry without disrupting the rest of the course.

Deep learning within a strategic frame. Through CIRTL I encountered the distinction between deep and strategic approaches to learning, which has shaped how I think about assessment. Students often start a course in deep mode and shift toward a strategic one as their workload accumulates. I have stopped resenting that shift. My goal now is to design summative work that lets deep learning survive inside a strategic frame. For the final term project, I built an analysis assignment around multiple simulations from an aerosol box model, asking students to trace how aerosol populations evolve under different process assumptions and different climate scenarios. They integrate size distributions, microphysics, and radiative effects into a single deliverable they can plan their time around. In an end-of-unit survey, 80% of students rated the active-learning approach more effective than traditional lecture, and 80% rated the real-world connections at the top of the scale.

I think about inclusion through the lens of the affective domain, not just the social and demographic background, but also the prior knowledge students bring and the level at which they engage with the material. My students in this course were predominantly chemical and environmental engineering undergraduates, with a small number of graduate students. They arrived with strong quantitative instincts but limited intuition for the interconnected, scale-spanning reasoning atmospheric science requires. When I introduced new material, I leaned on analogies from their own majors to have something familiar to attach to.

Mixed-level courses can quietly stratify, with graduate students dominating discussion and undergraduates going quiet. Anonymous polling and word cloud activities flatten this hierarchy enough that an undergraduate will commit to an answer a graduate student might later challenge, and will defend it because it is theirs. The egg-boiling activity does similar work for cultural and experiential background. The variation in answers reflects real differences in students' lives: where they grew up, whether they cook, whether they have ever lived at altitude. Once students see that their choices on the screen come from legitimate differences in their assumptions rather than from disinterest, disagreement becomes more diagnostic than adversarial.

My approach as an advisor is to make the goal of a lesson or assessment clear without removing the difficulty the student needs to work through. I state where we are heading, let the student take the next step, then meet again to debug what they tried. I treat one-on-one work as part of the same project as classroom teaching. During the term project, I held extended office hours to walk students through the project structure, debug their calculations, and discuss what their plots were telling them physically. Several students who described themselves as programming-averse or data-averse at the start produced full multi-panel analyses by the end. As I move toward an independent academic position, I expect to advise graduate students on aerosol-climate modeling and undergraduate students on shorter projects connecting box-model output to local air-quality questions.

I am interested in teaching existing undergraduate and graduate courses in atmospheric chemistry, air pollution, aerosol physics, atmospheric and climate modeling, and introductory climate science. Drawing on my research in aerosol-climate interactions and my co-teaching at the University of Arizona, I would be well-prepared to develop new courses on:

• Aerosols, Climate, and Health, integrating microphysics, radiative forcing, and exposure science for a mixed audience of engineering, environmental science, and public-health students.

• Computational Atmospheric Modeling, a project-based course using box and column models to teach numerical thinking through atmospheric problems.

• Climate Intervention and Geoengineering, a discussion-driven course that pairs the physics of stratospheric aerosol injection and marine cloud brightening with the Oxford Principles and current governance debates.

CIRTL taught me that teaching, over the long run, means becoming a reflective practitioner who keeps evaluating, reading, and revising. The most honest thing I can name from this semester is slide density. I tried to compress more onto each slide than my students could absorb in real time, and the survey reflected that. I was anxious about leaving anything important out, and packing the slide felt like the safest hedge against students missing something. What I did not yet have was the judgment to decide what to show and what to leave off. The next time I teach this material, I plan to cut slide text by at least half, lead each slide with a single question or image, and let what I say carry the work the bullet points used to.

One thing shifted over the semester. At the start I used the predict-then-explain sequence, a well-established active-learning technique I first encountered through CIRTL. In practice I found that stating a striking fact first and questioning students back into it held attention better and left a sharper takeaway. The move is consistent with the "time for telling" idea above, but it inverts the order: the fact is what creates the time, and the questioning fills it. Predict-then-explain is still a useful tool but starting with the fact has simply worked better for me.

The aerosol course I co-taught through CIRTL was a good test of the habits I care about, because none of the phenomena we studied have a single cause. Every interesting problem in atmospheric chemistry asks students to reason across scales, to carry uncertainty along as part of the answer, and to notice that the right question is often more useful than the fastest calculation. I plan to keep teaching this way where I can plant the few ideas I want students to keep long after the course, when the equations have faded.