Open to research collaborations, fellowships, and roles at the intersection of learning, data, and product.
English · Hindi · German
The front of this card has everything you need to reach me.
End-to-end redesign of a learning experience. Scoped learning objectives, restructured content flows, and built an evaluation framework to measure problem-solving outcome gains.
Evidence-based PD intervention giving teachers structured diagnostic protocols and feedback loops to support classroom problem-solving at scale.
PD program for 5th grade math teachers addressing student math anxiety — grounded in research on teacher efficacy and formative feedback design.
Empirical study comparing video instruction vs. learning-by-doing. Applied CTA to redesign the instructional sequence based on learner performance and cognitive load findings.
Raspberry Pi + OpenCV system that gamifies kinematics. Built physics tuning, a scoring model, and feedback loops — turning abstract physics into tactile play. Started in 11th grade out of frustration with a textbook.
Card game designed to facilitate structured dialogue among strangers. Tested at CMU — participants consistently formed stronger social ties through a single session than through standard interaction.
Study exploring pattern-seeking cognition and apophenia in teenagers using game-like prompts. Designed measurement instruments to observe behavior across repeated trials.
Play is a serious design tool. Samvaad worked because structured play lowers social stakes and raises genuine engagement. The Free-Throw Simulator worked because embodied feedback teaches physics better than a diagram. I use game mechanics where they earn their place — not as decoration.
You can't design learning without measuring it. The NLP pipeline, the CTA study, the Eberly observation protocol — these exist because intuition alone isn't enough. Good instructional design is testable. If I can't tell whether something worked, I haven't finished designing it.
The best interventions work within constraints. TeachingTeachers was designed for teachers with limited time. Unifly's diagnostic ran on a shoestring budget. Elegance, for me, means solving the actual problem with what actually exists — not the ideal version of either.
"Students deserve better systems, and building them is tractable."
Building an NLP evaluation pipeline to classify learner reflection depth and self-explanation quality in an AI-assisted Python tutoring study.
Supporting a redesigned math course for 30 students with repeated course failure histories; refining instructional sequences via scaffolding and retrieval practice.
Classroom observations across 200+ students; redesigning observation protocol to translate pedagogy notes into analyzable data.
Automated donor operations, cutting manual overhead by 30%; designed a custom school-readiness diagnostic tool.
Revamped internal training platform; reduced new-hire queries by 45% post-launch.
Thematic and quantitative analysis of multilingual STEM learning data from Yiya AirScience participants in low-resource environments.
Carnegie Mellon University — update course names as needed
Carnegie Mellon University · Expected May 2028
OR/Stats + Learning Engineering student building at the intersection of learning systems, educational design, and applied research.
I started sketching redesigns of my classroom — imagining what school could feel like if it were built around how students actually think. Didn't know the word instructional design yet. Just knew something wasn't working.
After staring at a lifeless physics textbook, I built a computer vision system to gamify kinematics. If the material wouldn't earn attention, I'd engineer something that would.
I designed social card games to spark genuine conversation among strangers. Watched friendships form in real time. Play, it turns out, is one of the most reliable ways to build real connection.
Courses, labs, and nonprofits gave me formal vocabulary for what I'd always been chasing: learning engineering, scaffolding theory, cognitive load, spaced reinforcement.
I work across data, design, product, and research. The throughline hasn't changed: students deserve better systems.
The throughline hasn't changed.