I chose it because it is the only piece where I can show my work being wrong and then getting better. The draft went to three classmates for review, and their feedback is documented alongside what I changed in response. Most portfolio pieces show a finished object. This one shows the process that produced it.

A twelve-page technical report on whether AI can help doctors catch cancer earlier than they currently do. It walks through the evidence, a 2026 trial where AI-supported mammogram reading found 29% more cancers, and colonoscopy studies where AI raised detection rates from around 33% to 41%, and then argues for a specific position: that these tools should support radiologists rather than replace them.

What it taught me is that I write long. One reviewer said my sentences in two sections leaned so heavily on semicolons and dashes that the text became heavy to read. She was right, and fixing it did not lose a single technical point, it just made the same argument easier to follow.

Structured written communication at report scale: letter of transmittal, executive summary for readers who stop early, argument body with IEEE references, limitations section for credibility, and Appendix A documenting peer feedback and revision. It demonstrates completeness, coherence, conciseness, and correctness under a formal engineering format.

The report is navigable at scale: generated table of contents, labelled sections, figure captions, and an executive summary for readers who stop early. Peer feedback is documented in Appendix A with the revisions that followed. One reviewer flagged heavy sentence density in two sections; I shortened those without dropping technical content. A reader can follow the full evidence chain or stop after the summary — and the revision trail shows the draft was tested on real readers.

I chose it because it is the most uncomfortable writing I did this term, and the constraint is what makes it worth showing. One page, no appendix, no figures, no room to recover a vague sentence. I had to sound credible about myself, map UniSpot and Undercut to a specific job posting, and close with a clear ask — all before the reader decides whether to keep reading past the first paragraph.

Assignment 1 was a one-page cover letter responding to a real IBM internship posting. The task was correspondence, not report writing: lead with what the hiring manager needs, not what I wanted to say first about my projects.

It forced the same audience judgement as the brochures, but without a second version to fall back on. Every sentence had to earn its place on a single page. I could not add a limitations section or an executive summary to rescue a weak opening.

The letter also exposed a Correctness gap I had not fully faced. It shipped with a missing article and a garbled closing clause even after I had done the proofreading exercises on someone else's text. Catching errors in your own work under deadline is a different skill from spotting them in a class handout.

Persuasive written correspondence under a hard page limit: credible self-presentation, mapping projects to employer needs, and a direct closing ask. Demonstrates Axiom 6 — communication involves interpersonal risk — and the 7 C's under pressure: Clarity and Conciseness when every line counts, Concreteness when claiming project outcomes, and Correctness when there is no appendix to fix a sloppy sentence.

The PDF fits the one-page constraint and is scannable in under a minute: opening states role and fit, body maps UniSpot and Undercut to the posting's requirements, closing makes a direct ask. Structure leads with the reader's question — why hire me — rather than project chronology. Proofreading misses noted in my reflection show where Correctness still needs work, but the letter communicates persuasively within correspondence norms.