Goals

My output now lives in too many places: repositories, a DOI-tagged release, conference materials, essays, performances. The undertaking is a personal site that gathers it into one legible home, with separate sections for projects, goals, music, a lookbook, and contact information, and each item able to open into its own page. The structure matters more than the first impression, because a finished project should snap in as a new entry rather than forcing a redesign, so the technical work includes a maintainable content system, reusable components, deployment through GitHub, and documentation for adding entries myself.

Part of the motivation is college applications, since people I contact will look me up. The longer-term part is a public record where anyone can trace a claim I make to the code or write-up behind it.

The activities list looked like a form and behaved like a compression problem: ten activities and five honors, each squeezed into position, organization, and description fields with hard caps of 50, 100, and 150 characters. I worked through every entry to pack verified numbers and strong verbs into the space, checking exact character counts rather than trusting estimates, and cutting anything already covered elsewhere in the application.

The second pass mattered as much as the first: an adversarial cross-review of the finished list against my master spreadsheet, hunting inconsistencies and overclaims. It caught real ones, including a phrasing that implied I established the statewide competition rather than founding my school’s program, which I corrected before it could cost me credibility with a careful reader.

The scalability argument from my research deserves audiences outside science fairs, so I ran a small outreach campaign with concrete asks. I drafted and sent a speaker pitch to a DFW climate organization requesting a short slot at a chapter meeting, applied to a youth climate program with a proposal for a piece on scalability in clean energy, and compiled a working list of platforms that genuinely accept community-submitted writing after weeding out the ones that only look open. The message stays constant across formats: moving clean energy from dream to reality, in language a general audience can act on.

For the UChicago extended essay I paired two songs that undo each other and decided the argument needed a form: a two-page designed file where the page turn itself performs the essay’s pivot. Page one runs a comic night-city aesthetic with an upside-down answer line at its base; page two flips to a cream record-sleeve world; a musical score rail crosses both pages and closes a complete D.S. al Coda structure.

The build was genuinely technical. I implemented it in HTML and CSS, rendered it to PDF with a headless browser, and iterated through four full versions: misregistered ink passes, kinetic typography, embossing effects, and a file-size rescue that cut the PDF by two-thirds by converting raster textures to vector. Feedback from an outside reviewer was triaged item by item, adopted, compromised, or rejected with reasons.

Ahead of my chemistry final I built the studying as a structured quiz cycle instead of rereading slides: work through each chapter’s lecture material and practice problems, then get drilled one question at a time with immediate correction. The topics ran from intermolecular forces, boiling point ranking, and heating-curve calculations through phase diagrams and into solution chemistry, electrolytes, dissolution energetics, and concentration units. The method caught real errors early, including an acid dissociation I had written splitting at the wrong bond. The goal was simple and specific: walk into the exam with every topic already tested at least once.

The Goldwater application forces one clean account of the research, and I built mine deliberately rather than by instinct. I studied how successful applications handle the prompt’s hard parts, working through a seminar walkthrough of a winning essay line by line to see how it braids career goals, coursework, and research into one argument.

My own materials followed: a research essay tracing the progression from scalable rGO to cleaner nanotube samples and stating what the D-parameter work shows and what it cannot yet show, a title settled after real deliberation so the novelty claim is precise, an impacts essay anchored on my father’s story, and a journal shortlist that makes the publication plan concrete instead of aspirational. After harsh outside criticism, I redid fits and rewrote sections rather than protecting the draft, and finished with a sentence-level claims audit.

Soon after moving to Denton I kept a standing appointment with a DDR machine at Free Play, the local arcade. The goal was unserious and completely real: get good enough to clear higher difficulties, which meant learning to read arrow streams ahead of the beat, building the stamina to survive a full set, and losing publicly a lot along the way. It was one of the ways a new city started feeling like mine, and it remains the only physics-adjacent activity I do where the only measurement that matters is the letter grade on the screen.

H2STAR is my self-directed study in adsorption thermodynamics, structured as a sequence of gates so I cannot skip the unglamorous parts. Early gates were exactly that: extract the modified Dubinin-Astakhov model and its fitted hydrogen parameters for a benchmark activated carbon from the primary literature, map the paper’s symbols onto the manual’s, populate a validated configuration file, and flag the paper’s own internal inconsistencies before trusting its numbers. Later gates digitize reference isotherms and build toward system-level storage estimates.

Two rules make it distinctive. Validation criteria are registered before results exist, so the goalposts cannot move, and every AI interaction the project uses is logged in structured form, so the final write-up can show exactly where assistance ended and my judgment began.

Being asked to keynote at my university, as its youngest undergraduate speaker in that slot, came with a specific communication problem: the audience would include students, families, faculty, and researchers, and the talk could not be a lab presentation with nicer slides. I built an argument instead, connecting hydrogen storage research to the energy demands of food systems, and developed it directly with university administrators so it fit the occasion.

The preparation was its own project: structuring the through-line, choosing which technical details and research moments earn their minutes, cutting accomplishments that did not serve the argument, and rehearsing pacing and transitions until the talk sounded spoken rather than read, for a room where most listeners would never have seen an XPS spectrum and should not need to.

My first full research paper turned the reduced-graphene-oxide experiment into an argument rather than a collection of spectra. Getting there meant developing the underlying protocol until it was reproducible: dispersing rGO in deionized water, depositing films on silicon-dioxide wafers, and defining irradiation conditions consistent enough that untreated and treated regions could be compared without preparation artifacts dominating the Raman response.

The paper itself needed a structured literature review, clear Raman-fitting methods, and a results section built around the D-prime to G separation and the D to D-prime intensity ratio. Writing it forced me to distinguish oxygen removal, increased sp3-type disorder, and actual hydrogen adsorption instead of treating them as interchangeable, and to explain honestly why preexisting edges and disorder limited the evidence. The submission had to stand on its analysis rather than an ambitious title.

With Team SMAT the objective was straightforward: work through as many picoCTF problems as we could solve correctly. The competition kept switching among cryptography, web exploitation, reverse engineering, forensics, and general programming, often with little warning about which concept mattered. My contribution was breaking problems into testable hypotheses, writing small scripts when manual work was too slow, and documenting solutions so teammates could build on them. None of it feeds my research profile, and that was part of the appeal. It was practice staying methodical when the system is unfamiliar and the intended path is deliberately obscured.

Presenting at the APS Global Physics Summit meant preparing a research talk for a room of working physicists: compressing the irradiation and characterization story into its allotted minutes, sharpening the distinction between oxygen removal and possible hydrogen-related sp3 defects, and treating the ambiguous parts of the data as evidence to present rather than details to hide. Compared with my first collegiate talk on the project, the bar moved: tighter definitions, less background physicists already know, and rehearsed answers about what Raman can and cannot establish.

The talk was recognized among the undergraduate oral presentations, but the durable value was different. Conference hallways are where the field actually talks, and a conversation there with a Penn postdoc about how composition changes reshape a crystal’s electrical behavior ended up threading through my thinking for months afterward.

I took over as Physics Committee head with one aim: make tutoring more useful than an open room with no plan. Sessions are organized around the problems students are actually facing in mechanics and electricity and magnetism, with short concept reviews, representative calculations, and practice sets that can be reused. The work includes recruiting tutors, choosing topics, preparing materials, advertising sessions, and adjusting explanations when different students are stuck for different reasons. The concrete outcome I am after is a repeatable support system that can run before major assessments and remain usable after I have left the course.

I decided the resume should be engineered, and it turned into a real software project: a LaTeX source with custom entry macros, a sidebar and main column balanced in paracol, and a build loop that compiles, renders pages to images, and inspects them for defects. Version after version, the work was concrete: hunting overfull boxes to zero, fixing widow lines, rebalancing sidebar spacing by the point, swapping macros so long date fields stop colliding with entry text, and verifying the page break lands exactly where intended. The other half of the goal was accuracy, checking every line against my master activities spreadsheet so the typography never outruns the truth.

As baritone section leader, the aim each week is to leave the section more secure than it started: check notes and entrances, isolate difficult passages, help younger singers with vowels and breath placement, and make sure our part supports rather than overwhelms the ensemble. The role is modest and concrete. Listen for where the section is losing confidence and fix that problem before the full choir has to stop for it. I also work on making the group comfortable asking for help, since a section sounds better when singers admit what they do not know early.

Storage only counts if the hydrogen releases, so the treated films go into temperature-programmed desorption, heated from 50 to 400 degrees Celsius while I watch the hydrogen and water-related mass channels for a desorption signal. The first campaign came back null across that window, which is itself information: either the binding is stronger than the window reaches, or the releasable coverage sits below the detection limit.

The follow-up treats reversibility as the real question. That means controls across pristine, annealed, and differently irradiated samples, plus repeated Raman and XPS after heating to see whether the bonding signatures relax toward their initial state or the treatment produced a permanent change. The aim is to state the release behavior of these films plainly instead of guessing at it.

The personal statement had one job: carry the Worldfest story, the slipped step, and what I learned about ambition sharing a stage with joy, in 650 words that sound like me and do the work no research essay in my file can do. I worked it paragraph by paragraph across many sessions, with specific craft targets rather than vague polish: open on the timestamp where the plan broke, fill the blank in the hook with a word that tells the truth without editorializing, keep the friend who pushed me into choreography in his own voice, and let the arc move from singular ambition to shared joy without announcing it. Every draft got scored against a rubric, and nothing stayed unless it earned the line.

STEMBox had three separate problems to solve, and I took the lead on all of them: design, money, and delivery. The design goal was three original physics kits that produce a clean, surprising effect a nine-year-old can operate, each inexpensive, safe, easy to reset, and documented with instructions and short video explanations so another student volunteer could teach it. The money goal was corporate sponsorships sufficient to build thirty units without charging schools anything. The delivery goal, carried out with volunteers through Protara Printing and JETS, became ten school visits and two science camps across Title I elementary campuses, reaching over four hundred students.

Watching the kits in actual classrooms reshaped what I think the product is. The hardware is the easy half; the sequencing and questions around it decide whether a demonstration turns into understanding, and that is where the next round of work lives.

At Hindi karaoke with my mom I have always joined in on melody and enthusiasm while faking half the verses. My training is Western classical, my ear knows these songs cold, and the missing piece is embarrassingly specific: the words. The plan is to learn a small set properly, lyrics and pronunciation, starting with Zaalima, so that joining in stops depending on someone else carrying the language. It is a modest goal with a clear test. Either I can sing a full song with her without the screen saving me, or I cannot yet.

Our transition radiation work was accepted for presentation at RADECS, the radiation-effects conference, in Prague, with me as first author. Acceptance created the real assignment: an eight-page proceedings paper written to IEEE standards, which will be my first publication of record and will face genuine peer review from radiation-effects specialists.

The work is line by line: a compact account of the six radiator designs, the beam setup, the detectors, and the analysis pipeline, figures that hold up under referee scrutiny, methods a stranger could follow, and claims sized exactly to what 441 million recorded triggers can support. A fuller journal version for IEEE Transactions on Nuclear Science sits behind it, so every choice in the proceedings draft has to survive a second, harder round later.

Joining the Judicial Board gave me a recurring responsibility rather than a single event to complete. The work is close reading: learn the Student Life Code of Conduct, go through incident reports and appeals carefully, and contribute to decisions that can be applied consistently across a residential community. That requires separating what is documented from what is assumed, asking whether procedures were followed, and taking both accountability and student context seriously. The standard is being reliable in the room, especially when a case is uncomfortable or the available information is incomplete, which means preparing before meetings and refusing to decide based on personality or reputation.

Before the nanotube work, my question was scalability: pristine graphene is impractical to produce in quantity, so can cheap graphene oxide and reduced graphene oxide powders undergo electron-irradiation chemistry consistent with hydrogenation, read through XPS and Raman? I developed that study into a complete science fair entry and took it as far as the circuit goes.

That meant the research itself plus everything the circuit demands: a poster that survives a judge standing two feet away, a research binder with the full record, forms and documentation in order, figures that show the result without hiding uncertainty, and rehearsed explanations at several levels of technical depth, including honest answers about controls, sample variation, and the difference between evidence of structural change and proof of stored hydrogen. Competing at ISEF in the materials physics category with my project partner put the work in front of judges from industry and research labs, exactly the audience a scalability argument needs.

The World Food Prize program asked for something my lab work never had: a written argument about food systems, defended live in front of experts. I wrote research papers connecting energy access to food security, investigating how hydrogen production, storage, and fuel-cell systems could support refrigeration, cold chains, irrigation, and rural energy needs, with country-specific evidence for Tanzania and Bolivia and attention to cost, infrastructure, and implementation. To explain the technology physically as well as on paper, I also built working fuel-cell demonstrations.

Carrying the argument through the program’s stages was the point, holding my own in roundtable discussions where professionals push back on student proposals in real time. It pulled my research into a policy-shaped room for the first time.

Comparing hydrogen-storage papers usually means an afternoon of unit conversion and charitable guessing, because uptake gets reported in mismatched units at unstated temperatures and pressures. HyCAN-DB is my answer: a screened, searchable database of carbon-based hydrogen storage results. The first phase was a systematic screen of candidate papers against a four-point relevance test, which passed twenty-eight of the first thirty examined.

The build phase is extraction and design: pulling uptake, temperature, pressure, and surface-area values into one consistent schema, normalizing units, grading evidence quality, and recording provenance for every number, aimed at an open release under FAIR data principles with a documented GitHub structure that later modeling work can query directly. The finished test is whether a stranger can use it and trust it without emailing me.

I approached Differential Equations section by section, and for each topic the standard was the same: identify the equation type, choose a method, carry the algebra cleanly, and check the result. The preparation ran through first-order equations, higher-order linear equations, series methods, transforms, and systems, worked through textbook problems and repeated practice rather than passive reading of solutions. The immediate objective was performing well on the exams, but the course mattered beyond the grade, because differential equations recur throughout mechanics, materials, and transport problems. I wanted the methods to become usable tools rather than isolated procedures I would forget after the class.

As committee head for Healing with Music, my job is the coordination that turns good intentions into actual visits: finding facility partners, lining up student singers and instrumentalists, confirming transportation and equipment, and shaping short flexible sets around music residents will recognize rather than music students want to show off. It is a different discipline from concert performance, because the audience shows its appreciation quietly and the measure of success is whether the visits keep happening on a rhythm instead of once as a feel-good event. The organizing goal is a program sturdy enough, in roster and routine, that it can repeat even during busy weeks and keep the music arriving for people who can no longer travel to it.

After the beam time ended, the experiment lived in the data: roughly 441 million recorded triggers spanning multiple beam energies, two silicon pixel detectors, and dozens of runs. As analysis lead I set out to determine which targets actually produced a transition-radiation signature, which meant learning the beam-telescope toolchain, cleaning the runs, matching target conditions, and working through occupancy maps, detector correlations, cluster-charge distributions, tracking, and residuals to separate possible soft X-ray effects from ordinary charged-particle backgrounds.

The central challenge was that the conclusion could not rest on one attractive plot. I needed a consistent analysis chain that explained why multilayer polyethylene and aluminized Mylar behaved differently from single-layer controls and the other targets, and that agreed with what the layer models predicted. That dataset became the backbone of everything the project published afterward.

Founding the chapter created the second job: teaching. I took on a cohort of twenty students with the aim of making them genuinely competitive in Number Sense, which meant teaching the shortcut systems one family at a time, writing practice materials, running timed rounds under real conditions, and proctoring the official contests. Part of the coaching was psychological, helping students grow less intimidated by a format where the clock is the main opponent. I kept competing myself alongside them, partly to stay sharp and partly because it is easier to coach a technique you still perform under pressure.

Five of them placed statewide. The number I care about more is twenty, because the design goal was that every student in the cohort, whatever their starting level, would walk into an official round prepared rather than hopeful.

VIPER, Penn’s dual-degree energy research program, became my early-decision target, and I treated the application as a project with its own workplan. The centerpiece is a long research essay written in lab-notebook format that walks a reader through the hydrogen storage work the way it actually happened, aimed at named faculty whose groups fit the problem. Around it sit the shorter supplements, each with its own lane so no two essays repeat material.

I also refused to write the fit argument blind. I set up conversations with the program’s managing director, a current VIPER student, and a professor whose research I cited, prepared briefing questions for each, and folded what I learned back into the essays. Finishing means a complete, internally consistent file where every claim survives a hostile read.

Raman conclusions in my field often rest on intensity ratios computed after a chain of quiet choices: despiking, baseline correction, lineshape selection among pseudo-Voigt and Breit-Wigner-Fano forms, peak set, intensity convention. My own graphene oxide analysis showed how easily those choices could move a scientific claim, so I built RamanUQ to make them visible, propagating them into honest uncertainty on the ratios through residual bootstrapping and model selection with information criteria.

I ran the project with unusual rigor for something written by one student: separate implementer and reviewer roles, a differential validation gate comparing my modules against independently written reference implementations on hundreds of randomized inputs, and a daily progress journal in my own voice. The release milestone was concrete: a versioned, documented package on GitHub with a Zenodo DOI, usable by someone other than me. The next milestone is adoption by a lab that is not mine.

Through Dull Roar I took on the unglamorous half of live music: booking, promotion, sponsor outreach, and run-of-show. The concrete goals were a series of charity concerts that actually raised money, which came to five shows, over a thousand dollars, and more than three hundred attendees, plus one event I was determined to see exist: a karaoke night big enough for over a hundred people, fully paid for by sponsors so nobody had to buy their way in. Running that night meant sign-ups, a moving queue, a tested audio setup, and an atmosphere where confident performers and nervous first-timers could both take the mic before exams closed in.

I also sang as a featured vocalist, which kept me honest about the logistics. A stage only works if the sound, the schedule, and the room were handled by someone, and I wanted to be the someone.

The D-parameter measurement deserves a paper, and I am writing it as first author with the claim sized honestly: the D-parameter tracks sp2 to sp3 hybridization in hydrogenated single-walled nanotubes, and it works as a hybridization probe rather than a direct hydrogen gauge, because no universal conversion from D-parameter to hydrogen weight percent exists in the literature. The draft needs a precise account of sample history, calibrated C KLL analysis set against C 1s and Raman, and a clean boundary between measured results and future work.

To position it I commissioned a publication-grade literature review built around three questions: how hydrogen content in samples like mine can be independently quantified, whether the new Raman feature near 150 inverse centimeters has precedent, and where nanotube hydrogenation research currently stands. An abstract version was accepted at a Materials Research Society national meeting, and my venue shortlist favors journals that referee for validity and move fast.

I agreed to choreograph a dance for Worldfest, our school’s livestreamed talent show, without ever having choreographed anything. The undertaking was concrete and unforgiving: pick the track, Tu Meri, break it into counts, invent movement that a group of mixed-experience dancers could learn, run rehearsals, and have it performance-ready by show night. It was also my way of stepping fully into South Asian cultural life at school instead of appreciating it from the audience.

We got it on stage. Partway through, my foot slipped and the plan collapsed, so the group finished the number on instinct while the stream kept rolling. We crowded into my room afterward to watch the video and laughed until we ran out of air. Underneath it all, I was proving to myself that I could take responsibility for a performance, teach something I had just learned, and deliver it in front of an audience.

Our team’s transition radiation proposal won CERN’s Beamline for Schools competition, selected as one of five from 508 teams across 72 countries. When beam time arrived at the ELSA accelerator in Bonn, I had just started early-college coursework and chose to participate remotely instead of missing weeks of classes. That decision defined the goal: contribute at full strength from another continent.

As design lead I built out the six multilayer radiator targets we tested, working with accessible films such as polyethylene, aluminized Mylar, and Kapton, choosing layer counts and spacings whose interfaces should produce a detectable soft X-ray signal, and accounting for frame absorption and geometry so the targets could survive installation and be compared fairly against single-layer controls. During the runs I coordinated remotely through my teammates’ accelerator shifts, and we presented results at CERN’s beam telescope and test beam workshop.

TAMS is full of people who love math, and when I arrived the main competition outlet was the AMC, which has a steep entry barrier and one narrow style of problem. Number Sense is different: eighty mental-math problems in ten minutes, no calculator, built on arithmetic shortcuts anyone can learn. Bringing the statewide Best of Texas circuit to campus meant coordinating with the competition directors, chartering the school’s first team, recruiting students, explaining the formats, and hosting official rounds on a regular schedule.

The chapter grew past sixty students, and I ran nineteen official contests while keeping the entry rule I founded it on: you sign up because you love math, and the skill comes afterward. The remaining work is the part founders usually skip, writing the operations down and training officers so the rounds keep running after I leave.

The experiment I planned in the Perez lab has a long chain of steps that can each quietly ruin the result: disperse purified HiPco single-walled nanotubes in DMF, spin-coat them into films on silicon dioxide, anneal them, load them into an ultra-high-vacuum chamber, and expose them to an atomic-hydrogen beam. Then comes the reading: C 1s and C KLL Auger spectra by XPS, Raman before and after, and the D-parameter as the gauge of how far the carbon shifted from sp2 toward sp3 bonding, with pristine and annealed films as the baselines.

The measurements handed back a puzzle worth the whole campaign. The D-parameter dropped substantially after exposure, the C 1s fit disagreed with it about the sp3 fraction, and a new low-frequency Raman feature appeared near 150 inverse centimeters. The goal became a defensible account of that set of facts: what the hydrogen bonded to, what each probe can and cannot claim, and how to present the D-parameter as a complementary measure instead of an exact hydrogen gauge.