YRI Fellows research almost anything they care about: AI, medicine, space, robotics, chemistry, climate, economics and more, each matched with a PhD mentor who has published in that area. Below are example topics from papers YRI Fellows have published or had accepted, organized by field and explained in plain English.

If your child is curious but has no idea what to research yet, that is normal. Most Fellows arrive without a topic. The PhD mentor helps them find one, and the first two weeks of the YRI Fellowship are spent finding the right question.

For the full list of papers, see Published Work. For the stories behind them, see Student Stories.

How YRI Fellows choose a research topic

A strong topic starts with what the student already finds interesting. YRI matches each student with a PhD mentor in their specific area, from AI and medicine to physics, chemistry, economics and psychology. The mentor helps shape that interest into a question that is specific enough to answer in 10 weeks and original enough to publish.

Some Fellows came in with a starting point:

  • Ruthwik Dhama wanted to study whether one exoplanet, K2-18b, could support life. With his mentor, that grew into METHI, a machine learning index that scores all exoplanets. It was published at an IEEE conference.
  • Nidhi Kulkarni had two separate interests. Her mentor helped her combine them into one project on detecting autoimmune disease from bacteria on the skin, published at IEEE ICCCA 2025.
  • Elliot Wong had sketched a custom-fit helmet idea on his own. It became an MRI-guided method for designing personal helmet liners, accepted at BioMRH 2026 in Glasgow, with proceedings in Springer.
  • Ben Zhang was a basketball fan who wanted to know why ACL tears end so many careers. He built models that predict ACL injury risk and recovery time from athlete training data.

Others came in with only a field they liked. That works too. Here is what Fellows have researched in each one.

AI and computing

AI is a common starting point, and it reaches into almost every other field. YRI Fellows have used it to make AI systems safer, faster and more reliable.

  • Stopping AI assistants from leaking private data. Rishik Asati, a 9th grader, built a defense that stops hidden instructions in emails and websites from tricking AI agents into leaking secrets. Across 150 test cases, active detection cut leaks of protected secrets to zero. Read Rishik's story.
  • Catching failures in huge cloud systems before they happen. Aditya Singla built a two-step neural network that detected 16 of 25 problems in cloud data, against 6 of 25 in earlier research. He won Best Oral Presentation at IEEE AAIML 2026 in Tokyo. Read Aditya's story.
  • Making AI reliable on a classic hard math problem. Manomay Jain showed that raw neural network answers to a famous optimization problem broke the rules in most cases, then built a repair method that made 100% of answers valid.
  • Finding a hidden flaw in physics-informed AI. Nishanth Alampally found and named the Jitter Paradox: the AI models that solve equations most accurately can be worse at rediscovering the physics behind them.
  • Teaching small AI models to track what they believe. Svanika Doddavarapu studied whether compact language models reason better when they track their thinking step by step.
  • Making AI-planned drone teams fail safely. Mohammed Alnuwaiser studied where safety checks belong when AI plans missions for groups of drones.
  • Predicting traffic accidents from a single camera image. Aarav Brahmbhatt, a 9th grader, built a model that predicts accident risk, waiting time and emissions from one traffic photo. It was published in IEEE Xplore.

Medicine and biology

Medicine is the largest field on the YRI record. Many Fellows combine biology with AI to tackle diseases that affect millions of people.

Brain and mind

  • Linking heart signals to depression. Devansh Gupta became a first author at an IEEE conference at 14 with a framework that uses ECG heart signals as an objective marker for Major Depressive Disorder. Read Devansh's story.
  • Spotting Alzheimer's from eye movements. Suriya Dev Saravanakumar, a 9th grader in Qatar, was accepted at IEEE EMBC 2026 in Toronto. Read Suriya's story.
  • Using astronomy techniques to spot Alzheimer's in brain signals. Ayaan Rustagi borrowed signal methods used to find faint signals in space and reached 88.9% accuracy detecting Alzheimer's patterns. He also won 3rd place at his science fair.
  • Reading movement intentions straight from brain waves. Aditya Arvind built models that read imagined hand movements from EEG, for patients whose muscle signals cannot be used to control a prosthetic.
  • A brain-inspired model of how we form memories. Ashwajit Warwatkar built a simulated network of neurons showing how replaying memories helps the brain hold on to them. It was published in Springer and presented at BICA 2026 in Mexico, two months after he enrolled.

Cancer

  • Teaching AI to know when its cancer predictions might be wrong. Akhilesh Kuppili tested how well AI predicts which melanoma patients will respond to immunotherapy, and whether it can flag its own unreliable calls.
  • AI that predicts how cancers respond to drugs. Arfa Haseeb combined several layers of biological data to help match patients with the treatments most likely to work.
  • Using machine learning to find lung cancer biomarkers. Aarnav Bhat built a model with an AUC of 0.983, a near-perfect score, accepted at IEEE ICITSIF 2026.
  • Designing peptide drugs for brain cancer. Aly Dhedhi, a 10th grader, designed and tested four candidate peptides on the computer for glioblastoma, accepted at IEEE CIBCB 2026.

Heart, blood and everyday health

  • Estimating blood sugar without a needle. Devansh Deka turned the light-based pulse signal from a simple sensor into images an AI can read to estimate blood glucose.
  • An explainable tool that screens for anemia from a blood test. Melissa Nnam, a 9th grader, built a tool that shows doctors which blood measures drive each prediction.
  • Decoding the gene variants behind a hidden heart disease. Ajwa Awan studied gene changes linked to an inherited heart condition that can cause sudden death in young people.
  • AI that finds gene-editing targets for heart disease. Atharv Ved built a CRISPR target prediction pipeline accepted at two IEEE conferences.
  • Flagging endometriosis risk without surgery. Harleen Badwal built a model that estimates risk from symptoms and history alone, reaching an AUC of 0.990.
  • A low-cost AI screening tool for rural clinics. Abhigna Naik built ArogyaScan, which runs offline and costs 70 to 80% less than a standard hospital ultrasound setup.

Genes and microbes

  • Tracking how Omicron mutated to escape immunity. Sai Pasuparthi identified 132 immune escape mutations across Omicron lineages.
  • Software to study disease in sea stars from DNA data. Sidharth Iyer turned independent research on sea star disease into his first accepted paper.
  • Finding the genes that help rice survive salty soil. Tarek Osama, from Egypt, found 28 strong candidate genes for salt tolerance in rice.

Physics and space

Fellows who love space have studied planets, satellites, pulsars and the signals that come from them.

  • AI that ranks which exoplanets could support life. Ruthwik Dhama's METHI index identifies the top 10 habitable exoplanet candidates, and he built a public website where anyone can look up a planet's score. Read Ruthwik's story.
  • Classifying gravitational-wave signals with free computing tools. Mubashir Suhail trained a lightweight neural network on free Google Colab and qualified to represent Pakistan at Regeneron ISEF 2026. Read Mubashir's story.
  • Predicting satellite collision risk before it escalates. Krishik Singh, a 9th grader, studied 10,460 satellite encounter events and showed that 25% of the eventually dangerous ones look safe early on.
  • How well telescopes catch the shortest exoplanet transits. Ana Perez measured how often short planet signals are missed depending on how often a telescope takes readings.
  • Can two nearby pulsars explain a cosmic-ray mystery? Yuvan Ghanta built a model to test whether two pulsars could explain extra high-energy particles detected in space.
  • Modeling what breaks light-based quantum computers. Raaghav Gundabolu modeled how losses in photons, channels and detectors limit photonic quantum circuits.
  • AI-designed heat shields for spacecraft. Arjun Vinayaka Doss used NASA materials data to find the best heat shield materials for reusable spacecraft.

Engineering and robotics

Engineering Fellows build things: arms, drones, helmets and machines.

  • A low-cost robotic arm you control with your muscles. Raeyaan Muppaneni built a bionic arm that reads muscle signals from the forearm and moves in response, with 93% accuracy on a $35 Raspberry Pi. It was published and presented at IEEE WcCST-2026. Read Raeyaan's story.
  • A drone that picks its own safe emergency landing. Amaay Gupta built Skyshield, which chooses one of seven emergency actions when a drone is in trouble, tested across 4.2 million simulated decisions.
  • Custom helmet liners designed from MRI scans. Elliot Wong, a 10th grader, designed helmet liners shaped to each person's head to help prevent injury.
  • A digital twin that makes factory machining smarter. Adit Agarwal built a virtual model of a machining process that adjusts to each part.
  • Turning everyday vibrations into electricity. Ariham Jain compared three designs that harvest energy from vibrations to power small sensors.
  • AI-optimized skin grafts. Ryan Vo started with no coding experience, learned Python and built an AI model to design better 3D-printed skin grafts, accepted at IEEE ICIEA26.

Chemistry and environment

These Fellows used science to protect water, soil and the climate.

  • An automated system to clean polluted rivers. Sai Tarekar, a 9th grader with zero coding experience, designed a system to clean the Yamuna River in Delhi. It was accepted at IEEE ICSCSS 2026. Read Sai's story.
  • Turning waste tires into cleaner fuel. Ishaan Menon designed a new catalyst system for breaking down tires into fuel. He published at an IEEE conference and became a state science fair finalist.
  • AI that picks materials to clean heavy metals from water. Naitik Gupta built a dataset of 289 experiments from 49 studies and an AI model that predicts how well each material captures toxic metals.
  • Mapping where weeds will grow before they appear. Arham Sethi built a weed risk system checked against field images and soil samples, with farmers of 20 to 30 years experience. It was accepted at IEEE CIBCB 2026 in Greece.
  • A climate action toolkit for schools. Aarush Mahajan built a toolkit for measuring carbon footprints and organizing community action, and placed 3rd globally in the Project Green Challenge 2025, as 1 of 14 finalists from more than 8,000 participants.

Economics, finance and behavior

Research is not only for labs. YRI Fellows have studied markets, money and how people make choices.

  • How sudden world events move India's stock market. Karan Aggarwal used a graduate-level economics method to study how two geopolitical events moved aviation stocks. His full paper was accepted at the World Finance & Banking Symposium 2026 in Riga, Latvia. Read Karan's story.
  • Fair, private AI for judging financial risk. Akul Nehra built an AI system for loan risk that protects privacy, treats groups fairly and meets three major regulatory standards at once. It was accepted at IEEE RCSM 2025, when he was 15.
  • Finding which people a message actually persuades. Saanvi Kumar built a model that predicts how each person will respond to a message, so a campaign can choose the best approach for each one.
  • Using AI to audit the research on GLP-1 drugs and binge eating. Athena Richter built a language-processing tool that sorts medical studies the way expert reviewers do.

Humanities and projects that mix subjects

YRI also matches students with mentors in the humanities: history, philosophy, ethics, and projects that mix subjects. A student who loves history or philosophy does not need to switch to science to do research. Many of the strongest projects on the YRI record cross fields, like Devansh's work joining cardiology, neuroscience, immunology and machine learning, or Ayaan's use of astronomy methods for brain disease.

If your child is not sure which field fits, the free YRI Global Research Proposal Competition 2027 accepts any subject, including historical analysis, economic modeling and philosophical inquiry.

Tips for picking your own YRI research topic

  1. Start with what you already wonder about. Ben's question came from basketball. Elliot's came from a helmet sketch. Sai's came from a polluted river.
  2. Make it smaller. Ruthwik started with one planet. A focused question is easier to answer and easier to publish.
  3. Combine two interests. Nidhi joined two ideas into one project. Cross-field work often finds questions nobody has asked.
  4. Do not worry about skills yet. Ryan and Sai both started with no coding. The mentor teaches what the project needs.
  5. Let the mentor help. Every Fellow is matched with a PhD who knows the field, and the first two weeks are for finding the question.

For more ideas, read how to choose a research topic, browse research question examples, or see how YRI matches mentors to topics.

What every YRI Fellow leaves with

Whatever the field, the YRI Fellowship works the same way. Your child gets a PhD mentor, 1:1, every week for 10 weeks, plus an editing team and a publishing team that submits to IEEE, Springer and Elsevier. If the journal asks for changes, we keep helping, free, until the paper is published. 87% of Fellows publish, and 73% place at competitions.

The Fellowship is $2,997, or 3 payments of $999, with scholarships decided case by case. Applying takes about 10 minutes, and you get a decision within 48 hours. See Tuition for details.

Find your topic with a PhD mentor: see the YRI Fellowship

Sources

SHARE THIS ARTICLE
XLINKEDINEMAIL