From market research to an evidence-backed resume

The research and compact evidence packet show a second use of AI: organizing role requirements and resume drafts around evidence. The packet separates established claims, items needing reconciliation, implemented independent work and designs still in progress.

The method stayed deliberately simple. Dated role requirements were compared with source evidence, and draft positioning emphasized the strongest supported matches. Gaps remained visible instead of becoming stronger claims during copy editing. A human kept every wording decision; the AI assisted with requirement synthesis and gap-finding, but it could not convert a missing artifact into a fact.

An illustrative mapping

For a role asking about durable agent state, an independent-project section can point to NIS checkpoint code and its bounded tests. It cannot turn that personal implementation into employer or customer deployment. This example illustrates the evidence rule; it is not a reconstructed claim about one particular application.

Linear Job Search is used for Career opportunity tracking. That operational use does not establish that Linear managed NIS implementation.

Keep proposals separate from completed work

Several boundaries were fixed in advance. A planned review chain — a Kimi-produced draft passed to a Nemotron audit — remains an unchecked TODO; separate architecture reviews do not prove a resume audit happened. The original vendor list and prices behind the historic market scan were never recovered, so no retroactive vendor bakeoff was reconstructed. Statements about the market carried their dates and were not treated as current recommendations. Progress was judged with a role-fit rubric — how well each claim matched stated requirements — never as a hiring probability, an ATS score, or a claim of time saved or business impact. No hiring outcome is attributed to this workflow.

What appears here is the sanitized public process; the underlying evidence register stays private, and no private resume, employer or recruiter details are reproduced here. The architecture story explains the system design, and AI-assisted implementation and evidence documents the coding side and its evidence trail.

The reusable lesson is evidence first, wording second. Map what you can prove to what a role actually asks, mark honestly what you cannot, and let no tool inflate the gap.