Research
From Research to Product
Initial Question
The first three research documents helped answer one question:
Can physiology explain why hardworking students sometimes fail under pressure?
By the end of that journey, we had reached a working model. Academic performance was not simply a function of knowledge. It was also influenced by the physiological state in which that knowledge was accessed.
The next question naturally became:
If this understanding is correct, how should educational products change?
Rather than continuing to search for theoretical answers, we decided to explore that question through engineering.
Working Model
The conceptual model that guided every engineering decision is shown below.
flowchart TD
A[Sleep] --> C[Physiological State<br/>Focus Score]
B[HRV] --> C
C --> D[Brain State<br/>Attention<br/>Working Memory<br/>Encoding<br/>Memory Consolidation]
D --> E[Academic Execution]
E --> F[Academic Performance]
Our assumption was intentionally simple.
Students rarely fail because knowledge suddenly disappears.
More often, stress changes the brain’s ability to retrieve and execute that knowledge at the required moment.
Knowledge remains.
Execution deteriorates.
That distinction became one of the foundational assumptions behind our work.
A Shift in Perspective
Most educational technology products optimize only one variable.
flowchart TD
A[Learning] --> B[Questions]
B --> C[Marks]
The implicit assumption is straightforward.
If students consume more educational content, they should achieve better academic outcomes.
Our research gradually suggested a different perspective.
flowchart TD
A[Knowledge] --> C[Brain State]
B[Physiology] --> C
C --> D[Execution]
D --> E[Marks]
In this model, improving academic performance is not only about increasing knowledge. It is also about preserving the physiological conditions required to effectively use that knowledge.
From Questions to Engineering
Once this working model became clear, several engineering questions naturally emerged.
Can physiology be measured continuously?
This question led to experimenting with wearable devices and eventually reverse engineering an existing smart ring platform instead of attempting to manufacture custom hardware from scratch.
The objective was not to build a wearable device.
The objective was to obtain continuous physiological signals suitable for future experimentation.
Can physiology be translated into a useful educational metric?
Raw HRV values or sleep duration are difficult for students to interpret.
This motivated the development of an experimental Focus Score—a higher-level representation intended to summarize physiological readiness in a way that could eventually integrate with academic decision making.
The exact formulation remains an internal research artifact.
Can interventions become personalized?
Traditional educational platforms deliver identical recommendations regardless of a student’s physiological condition.
Our working hypothesis suggested the opposite.
Two students with identical academic performance may require completely different interventions depending on their physiological state.
This motivated the development of Shiv, an intelligence layer designed to combine academic context with individual student context when generating recommendations.
Can interventions happen before failure?
Most educational products respond after poor performance.
Our research suggested that physiological indicators often begin changing before noticeable academic decline occurs.
If those signals can be detected early enough, interventions may become preventive rather than reactive.
This became another guiding principle for future product development.
Influence on Learnzy
The research did not directly produce a finished product.
Instead, it shaped the engineering direction behind several independent systems.
| Research Question | Engineering Direction |
|---|---|
| Can physiology be measured continuously? | Wearable integration and smart ring experimentation |
| Can physiological readiness be summarized? | Experimental Focus Score |
| Can interventions become personalized? | Shiv intelligence layer |
| Can stress be detected before performance drops? | Continuous physiological monitoring pipeline |
| Can learning adapt to the student’s current state? | Physiology-aware planning and recommendations |
Each engineering project explored one part of the larger hypothesis rather than attempting to solve the entire problem at once.
Relation to Existing Research
During this investigation, we found that several recent publications were beginning to explore similar directions.
Rather than viewing AI purely as a tutoring system, newer work increasingly combines:
- physiological sensing,
- stress estimation,
- sleep analysis,
- personalized interventions,
- adaptive learning systems.
Discovering this literature was an important validation.
It suggested that the broader research community was beginning to ask many of the same questions we had independently arrived at through our own observations.
The objective was never to reproduce existing work. Instead, these studies helped refine our thinking and increased confidence that the overall direction was worth pursuing.
Current Position
At this stage, we do not claim that physiology alone explains academic performance.
Academic success is influenced by numerous interacting variables including knowledge, motivation, environment, socioeconomic factors, teaching quality, and mental health.
Instead, our position is more modest.
We believe physiology represents an important variable that has received comparatively little attention within educational technology.
Rather than treating this as a proven fact, we treat it as an engineering hypothesis worth building around.
What This Research Changed
Perhaps the biggest outcome of this research was not a product or an algorithm.
It was a different way of thinking.
We stopped asking:
“How can students study more?”
and started asking:
“What physiological state allows students to perform at their best?”
That shift influenced every major engineering decision that followed—from wearable experimentation and physiological signal processing to personalized AI assistance and future intervention systems.
This document marks the transition from research into engineering. The projects documented in the Engineering section of this portfolio are practical explorations of the ideas introduced throughout this research journey, rather than final answers to the questions themselves.