§ Suite · Services

What Illuma actually does.

Live previewClick to collapse
What the AO saw
G12CS + Math3.94 UWSAT 15208 AP9 activities6 honorsessay 640w
Headline
A committed math student with an unusually clear teaching instinct.

Would advance in first-round reading at match tier. At reach tier, the missing research signal is what a strong reader would circle first.

1
Academics + rigor
Stats, GPA, course rigor, school context

3.94 UW / 1520 SAT is a clean academic floor. Course selection (BC Calc + AP Physics C + Multivar) reads as genuinely rigorous for a CS applicant — not a generalist filling seats.

What lands

AP Physics C 5 + Multivar 5 signal quant depth beyond just the CS curriculum.

What worries

No AP Chem or AP Bio — closes some pre-med hedge options, though that's not the target track.

2
Activities + honors
Impact, spike, passion signal

The math-mentorship arc is your real spike. AIME + USACO Silver + 30-kid mentorship program cohere as a real through-line — not a resume of separate things.

What lands

Founded mentorship (30 kids) is doing-the-work, not resume-padding.

What worries

No research or independent project — closes off top CS programs where publications sort admits.

Piano dilutes — 8 years with no external validation reads as filler.

3
Essays
Authenticity, story consistency, class-fit

Common App is thin (640w) — the mentorship story is there but the specific 'aha' moment isn't landed. Supplements missing for MIT + CMU.

What worries

Common App opener leans on a competition anecdote — safer to lead with a mentorship moment given the file.

What would sharpen this

No supplement drafts saved for MIT or CMU — school-fit read will be shallow until logged.

4
Recommendation letters
Validation from teachers + counselor

Rec letters aren't stored in Illuma — this phase runs blind.

What would sharpen this

Ask your math teacher to speak to the mentorship arc — the strongest signal in the file isn't yet backed by an adult voice.

Output includes
  • Per-school scope — reframes as that specific college's AO
  • Overall scope — portfolio-coherence view across every essay
  • Structured output: headline · strengths · concerns · coherence · major fit · verdict · data gaps
  • Feeds on your saved essays automatically — no re-pasting
  • Separates 'missing data' from real 'concerns' so we don't roast you for what isn't saved
Under the hood

ML models + calibration layers, not a ChatGPT wrapper.

01

School-scoped context

When you pick a college, the prompt reframes as that school's AO — and the file bundle drops other schools' supps because a real AO doesn't see them.

02

Portfolio coherence

Overall mode includes a bulleted map of every drafted essay so the read can flag drift or recycled anchors across the wider portfolio.

03

Data-gap discipline

The model is instructed to route anything missing (unsaved essay, blank honor) into a data_gaps list, not to critique you for what we don't have on file.