A small tool with better questions
N.S.’s small pronunciation tool became more persuasive when it stopped pretending to be an education platform and began revealing what he learned from real users who did not behave as expected.
Strong coding, fewer than thirty users, too much platform language
N.S. had built a simple tool that let younger cousins practice pronunciation in two languages. Early drafts called it an education platform even though it had fewer than thirty users and had never been formally evaluated. His academics were strong, but the application reduced him to grades, coding competitions and technology terminology. His interest in language had come from family life and translation, yet it appeared nowhere outside one line in the activities section.
His parents preferred a direct computer-science path and worried that discussing linguistics would make his goals appear unfocused.
What the project could honestly carry
What the application already showed
- Strong quantitative and technical preparation
- Independent coding initiative
- Interest in user behavior beyond the code itself
What it did not yet answer
- What could the tool actually do?
- What had N.S. learned from fewer than thirty users?
- Was linguistics a genuine interest or a late label?
Users, language, and the limits of scale
The questions below did not supply an admissions formula. They determined what evidence needed to be checked, which claims needed limits, and what the applicant still had to decide.
Why do users ignore an instruction the designer considers obvious?
Users ignored instructions N.S. considered obvious because the design assumptions belonged to him, not automatically to the cousins using the tool.
How should software respond to accent, hesitation, and language switching?
Accents, hesitation, and language switching made the technical problem inseparable from questions about how people actually speak.
What can a small project prove?
A project with fewer than thirty users could show iteration and learning, but it could not establish broad educational impact.
User confusion became more valuable than user count.
The profile was reorganized around a more specific question: how software responds to the way real people speak, hesitate and switch languages. N.S. documented what the pronunciation tool could and could not do. He described the number of users honestly, removed claims of demonstrated learning impact and included the feedback that exposed weaknesses in his design.
Essay sessions developed scenes from testing the tool with relatives who ignored instructions he considered obvious. This allowed him to discuss listening, usability and language without pretending the project was a startup. School research examined computer science, linguistics, independent work, breadth requirements and the feasibility of crossing departments. Rejection and waitlist scenarios were discussed before decisions so that one result would not define the entire cycle.
The facts stayed the same. Their hierarchy changed.
Evidence was made more precise, attributable, and useful. The goal was not to enlarge the record, but to stop one title, institution, hardship, or outcome from carrying more meaning than it could support.
The tool was called an education platform.
It was described as a small pronunciation tool with fewer than thirty users.
Impact was implied through ambition.
Capabilities, limits, and user feedback were stated directly.
Linguistics looked like a distraction from employability.
Program structures showed how it could deepen the technical interest.
An international choice that also had to make sense at home
N.S. was surrounded by examples of student founders whose projects were described through downloads, revenue and media attention. His pronunciation tool had none of those signals. At first he believed the only alternatives were to exaggerate it or omit it. A closer review showed that its value lay in the questions it exposed: why relatives ignored a prompt, how accents challenged the model and which corrections felt discouraging rather than useful.
The family conversation required care because his parents' preference for computer science was not simply about prestige. They were planning a large international expense and wanted a recognizable employment path. Instead of dismissing that concern, N.S. researched degree structures and career outcomes while explaining that linguistics strengthened rather than displaced his technical interest.
Independence remained visible in the work.
- N.S. retained control of the code, examples and writing. The mentor asked for evidence whenever a claim moved from description to impact. His parents received a clear comparison of academic structures, but N.S. made the final preference order.
Choosing room for computing and linguistics together
N.S. was admitted to Princeton, Berkeley and Cornell and denied by MIT and Stanford. He selected Princeton for the combination of independent academic work, computer science and access to linguistics. The two denials were recorded as normal results in an uncertain process, not explained away.
WHAT CHANGED
- The phrase education platform was replaced by an accurate description of a small tool.
- Capabilities, user count, and limits were stated directly.
- Family language experience became visible beside coding credentials.
- The family examined how linguistics could deepen rather than displace the technical path.
WHAT DID NOT CHANGE
- The tool had fewer than thirty users.
- It had not been formally evaluated for learning impact.
- N.S.'s strongest formal preparation remained technical.
- His parents still needed a credible academic and employment structure for an expensive international degree.
The reader’s understanding changed in stages.
This sequence describes what the revised evidence made easier to understand. It does not claim to reconstruct an admissions committee’s private deliberations.
A strong coding and testing applicant whose humanities evidence appeared thin.
A small pronunciation tool described in language too large for its evidence.
A designer learning from accents, ignored prompts, and discouraging corrections.
A student selecting a program where computing and linguistics could meet through real questions.
The alternatives were plausible—and less useful.
Keep education-platform language
The project would imply reach, evaluation, and maturity that it did not have.
Omit the project because the user count was small
N.S. would lose the clearest evidence of how real users changed his technical thinking.
Add linguistics as a label only
The interdisciplinary interest would sound strategic rather than rooted in behavior.
Dismiss parental career concerns
A legitimate question about international cost and degree structure would remain unanswered.
Each stage used a different test.
| Decision | How it was tested |
|---|---|
| How to name the project | Use a functional description and state the user count, testing, and limits. |
| How to connect disciplines | Trace the connection through accents, hesitation, instructions, and user feedback. |
| How to discuss impact | Claim only iteration and learning that direct observation supports. |
| How to choose a program | Compare computing, linguistics, independent work, and cross-department access in the actual curriculum. |
WHAT THIS CASE SUPPORTS
- N.S. built and tested a small pronunciation tool with relatives.
- User behavior changed the questions he asked about the design.
- His interest in language was grounded in family and project experience.
WHAT IT CANNOT PROVE
- That the tool improved language learning.
- That fewer than thirty users represented a market or broad population.
- That the project was a startup or mature education platform.
Questions before calling a project a platform
The profile shows how one applicant’s evidence and decisions were organized. It does not predict another person’s result or supply a story to copy.
- What did a user do that your design did not anticipate?
- Which feature works technically but poorly for the person using it?
- What is the narrowest claim your evidence can support?
- Would the academic connection remain if the project name disappeared?
- Can your family trace the degree structure without relying on a broad career label?
N.S.’s project became stronger when its limits were visible, because those limits revealed the questions he was genuinely prepared to pursue. A small project does not need large user numbers to generate worthwhile questions. N.S.'s application improved when inflated platform language was removed and user confusion became evidence of learning. Interdisciplinary interest was grounded in actual behavior rather than a last-minute academic label.