Two months ago, two Lakesider seniors each took home two thousand dollars and more importantly, the honor of being among the nation’s top 300 young scientists.
Aanika T. ’25 and Rohan D. ’25 were named Top 300 Regeneron Science Talent Search (Regeneron) Scholars early January for their submissions of original research to the “oldest and most prestigious science research competition for high school students,” claims the competition’s website.
Regeneron is run by Society for Science, a nonprofit encouraging scientific research. Applicants submit an overview on their unique individual research in conjunction with various essays.
Ironically, the goal of both seniors wasn’t to win Regeneron — for Rohan, it was to get his paper published in a scientific journal.
The goal started with USA-PRIMES (PRIMES), a Massachusetts Institute of Technology (MIT) program for high school mathematics research. Accepted students are paired with a mentor, usually a graduate student or postdoc with experience in their topic, and spend about a year exploring the cutting edge of mathematics, presenting their findings in a conference, and eventually publishing a paper.
With that paper, Rohan decided to enter Regeneron. His field of study in mathematics was knot theory, which explores the mathematical properties of knots. It’s usually applied in macronutrient research and quantum physics.
To break down the math lingo in his winning paper’s title, “Patterns in the Stable SL(N) Homologies of Torus Knots,” Rohan explains that a torus is “basically a donut,” and a torus knot is a string that wraps around said donut.
The meat, or dough, of his research were two main conjectures — unproven but likely theories — surrounding similarities in the behavior of knots. One conjecture, which he dubbed the PIE (Principle of Inclusion and Exclusion) conjecture, predicts the structure of said knots as they wrap around the donut arbitrarily many times.
Despite the randomness of donuts and twists, his theories have a real, though niche, impact in the fields of elementary particle physics, quantum physics, and protein and DNA folding. On the macroscopic level, Rohan suggests that his research could help with identifying superconductor candidates — a technology that could revolutionize transportation and energy production.
After reading a list of other hyperspecific real-life applications with bizarre names (“triply graded Hilbert space [sic] of BPS states.”), Rohan pointedly stated that “I don’t really understand [the applications], and I’m the person who wrote them.”
Curiously enough, knot theory wasn’t his first choice. While selecting topics for his PRIMES application, Rohan wrote “machine learning to solve partial differential equations” as his first choice. Sadly, Rohan had to settle with his second choice — put in because “I was quite interested, I figured no one else was going to do it, and I wanted to be cool and different. And also I figured it might help me get in.”
Another punch line — Rohan thought he “was not exactly the person who I thought was going to get into PRIMES because they have a list of four criteria — they say it is recommended you meet at least one. I met zero of them, but I applied anyway and I got in. I just solved all the problems on the problem set, wrote 60 pages, and created an extension question of my own.”
After attending a welcome meeting on January 29th, he met his mentor in early February: Dr. Joshua Wang, “chill” and “fun” MIT postdoctorate, and Princeton Veblen Research Instructor.
“He’s an absolutely fantastic teacher, brilliant mathematician as well, and definitely knows what he’s doing […] This guy could teach me what would be a week of math for a graduate in an hour,” Rohan said.
Ultimately, despite his advisor’s educational aptitude, math isn’t done in a night: thus came two months and 80-100 dense pages of intense reading about topics like homological algebra.
After the learning phase, Rohan set out to create a conjecture by June. Though he ultimately came up with two, he admitted that “it turns out they were pretty hard to prove.”
However, Rohan remembers that “my mentor also made it pretty clear to me when I told him the conjecture; he was like ‘I’m going to be amazed if you can prove this because this is crazy.”. With the pressure of November to start drafting the paper, Dr. Wang worked with him to crack the mystery, starting with finding patterns in “simple versions of these complexes.”
“When I was bored in class, I’d start writing stuff on a piece of paper. Not saying which class — the wolves will be airdropped upon me,” remembers Rohan. Half of the time, he would be lying on his gray leather couch. The other half was spent at twin two by three foot whiteboards “furiously scribbling things” like a scene from the movies. In true mad scientist fashion, Rohan also made good use of his blackboard.
But with the 10 hour weekly commitment, and (not literal) wall headbanging from the complexity, one day, a “eureka” moment came — while drifting off, nonetheless. He quickly jotted down a few notes on his iPad. When he woke up, the realization stuck, and “ended up being a good part of the paper.”
Ultimately, four months of research and four months of writing culminated in a two-day weekend PRIMES conference at MIT from October 12-13. Landing Friday afternoon, Rohan immediately got to work… meeting Lakeside alumni. Friday night included eating at Maseeh Hall with recently graduated students, followed with residence meet-ups. Other nights were flipping coins into circles for various self-inflicted “punishments” or evening strolls around Cambridge.
However, MIT didn’t subsidize his trip just to dilly-dally with friends. Saturday morning, after nearly missing his presentation because of locked building fiascos, Rohan navigated his way into room 4-370 in the Maclaurin Buildings and delivered a 12-minute presentation followed by 10 minutes of questioning from students (mostly friends of PRIMES researchers, Rohan guesses) and curious professors.
The final step in completing his time at PRIMES was to compose a paper to publish in a magazine — Quantum Topology, specifically. Now, all that’s left is editing with his mentor to make the paper “readable” to mathematicians.
But Regeneron? Rohan claims he spent two days on his application.
“It was the middle of college app season, so I kind of forgot about it until the week before [the deadline] when I got an email. I was like, ‘Oh, I should probably do that,’” recalls Rohan. After wrapping up some college applications, just two days before the due date, he “copy pasted” and quickly drafted a “truly insane” amount of essays before turning it in.
Rohan was with grandparents in India early January preparing to sleep when his mom strolled into his room and alerted him of the news.
“I was slightly surprised because I did finish the application in two days, and [Regeneron] didn’t really have a propensity for math projects because they’re a little less applicable than, say, computational biology,” Rohan explains.
Nevertheless, he continued on with his life the next day. He reiterates that PRIMES wasn’t meant specifically for Regeneron — rather, as a student who’s always been interested in puzzles, it was to participate in “the only program that allows high school students to do real math research.” He adds: “math research is, in effect, a puzzle.”
“I mean, you look at the papers that PRIMES produces, and they’re regularly at the level of PhD papers. […] Just to be able to do something you theoretically shouldn’t be able to do for another six to eight years, and meet people from around the U.S. who are doing the same thing, and meet mentors who know what they’re doing; in that respect, it is one of the greatest parts of the PRIMES program.”
For Aanika, the other Lakeside winner, it was the pressure of the Eastside Indian community that led her down a long path of science fairs, and eventually to the international stage.
Her prize-winning research’s topic originated close to home. Since she had autistic cousins, Aanika remembers being very intrigued by neurodiversity. This was further piqued by her older sister, who was also interested in how neurodiverse brains worked.
Doing research on autism, Aanika stumbled upon a video called The Autism Broadcast at Seattle Children’s Hospital. That was how she met Dr. Sarah Jane Webb, a UW graduate and a developmental cognitive neuroscientist.
Thus came the challenge she wanted to address with her research. Dr. Webb explained to her the issue of underdiagnosis in female autism specifically. So after feeding data from questionnaires and interviews with a machine learning model, she discovered a discrepancy between male and female diagnoses.
Overall, after isolating the ten factors that her trained diagnosis model prioritized, Aanika discovered that female autism diagnosis was more accurate in questionnaire and interview based diagnosis compared to an hour-long observation period. To improve accuracy, Aanika proposed “modify[ing] diagnostic criteria specifically for females to emphasize questionnaires and input from caregivers.”
Additionally, she found 12 “predictive features of female autism,” narrowing it down from 23 prior benchmarks, satisfying her goal of “mak[ing] female autism diagnosis more efficient, successful, and accurate.” With a random forest model, she eventually trained an algorithm with 92% accuracy, up from the previous 85% accuracy of the old 23-feature model.
In terms of challenges, since her research was built upon hours and hours of reading, it felt like a standstill sometimes, even meaningless. Alas the looming deadline of science fairs (along with her passion for coding) kept her sifting at the mounds of data she located. Overall, the time commitment was around three to four hours a week.
But once again, this research was initially intended for science fairs. The seed was first planted thanks to family friends from Tesla STEM heavily involved in science fairs. Back in her junior year, Aanika was interested in computer science and health, they seemed like the perfect option.
The journey to the International Science and Engineering Fair (ISEF) all began regionally, so on a random Saturday afternoon in March, she hopped into a Zoom meeting and waited very nervously for her turn to present her Behavioral and Social Sciences project to a panel of 4 judges — all part of the virtual Central Sound Regional Science and Engineering Fair (CSRSEF).
Luckily, despite being very scared, the 8-10 minute pitch went “really, really well,” thus landing her a spot at the subsequent round among the 12 out of 240 entries. This time, she had to face ten judges, competing against other students who were at the top of their category too.
“It was so scary. I was really, really happy because [among] 200 plus projects, I was chosen to be a finalist. So I had a chance.”
The next Friday was the big night. On a zoom call, the judges handed categorial awards one by one. However, when another girl won a behavioral and special sciences award, a moment of uncertainty lingered. But soon enough, Aanika would find out she still won first in her category. And when the judges flipped to the slide of ISEF qualifiers, her face was there.
“When I found out, I was in shock. My entire family just started screaming. I started crying. I’m like, ‘I’m going to ISEF.’ It was just a crazy, crazy experience because again, a week before that I never participated in a science fair,” Aanika reminisces. After, a trip to grab Ben & Jerrys and Thai food was in store as congratulations.
Qualifying was only the beginning though. After the news in mid-March, it was about a month of intensive work constructing a physical poster and refining her pitch.
On her birthday, May 11, Aanika and the rest of Washington’s team flew to the Los Angeles Convention Center. That night, she had a little birthday celebration with her parents, featuring a red velvet cake and watching Anyone But You.
The next morning, upon entering the venue, Aanika recalls seeing massive banners looming from the ceiling. “It was insane. ISEF is truly a global competition; I saw more people from other countries than from the US.”
Overall, despite the diversity, Aanika regrets not meeting enough new people, citing the pin exchange. “Each country has pins, and you can go around meeting new people and exchanging pins. The Australians have like little koalas on their pins,” she explains.
The actual competition began Monday. There were about 50-75 projects per row, each with their own booth. Ironically, Aanika was “so chill” compared to her first-ever science fair. Overall, through eight hours of intense judging in fifteen minute blocks, she received twelve judges — two that swung by simply out of interest.
But her most impactful conversation wasn’t with an official but rather two girls with undiagnosed autism that approached her project, which Aanika remembers helped “validate a lot of their struggles. It’s going to sound cheesy; talking to people who are directly impacted is definitely the best part because they give me valuable insight. I see my purpose is rejuvenated.”
Other interesting projects included a girl from the Midwest who fed her classmates crickets to investigate a novel protein source in a place of higher food insecurity.
In the future, Aanika hopes to confirm the accuracy of her models with a researcher at George Washington University using their clinical dataset. In an ideal world, her research would impact neuroscientists across the world, starting with updating the Diagnostic and Statistical Manual of Mental Disorders, a widespread reference manual, to recognize the disparity between male and female autism diagnosis.
Ultimately, like Rohan, Regeneron wasn’t her main goal; rather, it was to explore research opportunities through science fairs. It was simply opportunities like these that opened up.
“The [opportunities] are just so incredible. I’m so honored and grateful to have experienced them in high school. I truly did not think anything would come from this — not in a pessimistic way. Competing with people who had so much more experience, so much more support; the biggest part of it for me is just that in a year I was able to do this. I’ve gained a lot of confidence.”
