Stephanie Luciano / portfolio
Updated May 2026
Built with Claude

A working portfolio of things I made with AI.

Ten-plus years in HR AdTech, strategic partnerships, and account management. The last six months rebuilding how I work, one Claude project at a time.

Approach

I've been learning AI the way I learn anything, by doing. These are four workflows I built with Claude over the last six months. The ones that actually stuck.

01

Daily Gmail brief

The problem

My personal inbox was a daily mix of school updates, recruiter outreach, appointment reminders, party invites, and noise from newsletters I'd forgotten I subscribed to. Reading it cold every morning was eating 30 minutes before the day even started.

What I built

A Cowork scheduled task that runs every weekday morning. It scans the last 24 hours of Gmail, sorts everything into five priority buckets (school comms, recruiter outreach, appointments, birthday invites, unsubscribe candidates), flags urgency per item, and drafts a brief to my own inbox.

See the prompt
Every weekday at 7 AM, scan my Gmail for messages received in the last 24 hours. Categorize each into one of five buckets: 1. School communications (Hola! Charter, teachers, IEP, school events) 2. Recruiter outreach (new role pitches, interview scheduling, follow-ups) 3. Appointments and reminders (doctors, deliveries, calendar confirmations) 4. Birthday parties and social invites (RSVPs needed, kid party invites) 5. Unsubscribe candidates (newsletters and promo emails I haven't opened in 30+ days) For each item, flag urgency (urgent / today / this week / fyi) and pull the most important detail. Email me the brief by 7:30 AM with subject line: "Inbox brief — [date]". Keep it scannable. No fluff.
Tools
Claude Desktop (Cowork) Gmail connector Scheduled tasks
Impact

30 minutes back every morning. I open my inbox already knowing what needs action versus what can wait, and I've cleared a noticeable chunk of low-value subscriptions.

02

Interview prep system

The problem

Late-stage interviews at the Director / Senior level need to feel like you already work there. Generic prep doesn't get you to top of round.

What I built

A repeatable workflow that turns a JD plus a recruiter screen transcript and hiring manager background into a tailored prep doc. Includes a company snapshot, role-to-experience mapping, a STAR-format story bank pulled from my actual career, smart questions to ask, and an honest read on what the hiring manager will probe hardest.

Used for

Four active senior partnership and account management interview processes spanning payments, grocery marketplace, job board, and SaaS partnerships.

See the prompt
I have an upcoming interview for [ROLE] at [COMPANY]. Interviewer: [NAME], [TITLE]. Build me a prep doc with these sections: 1. Company snapshot — recent funding, ARR if public, leadership, key product moves in the last 12 months, biggest competitor and what differentiates them. 2. Hiring manager background — career arc, what they've said publicly (LinkedIn, podcasts, blog), what their team likely looks like. 3. Role-to-experience map — top 5 responsibilities from the JD, each mapped to a specific example from my career. Use my resume and past chats as source material. Be honest about gaps. 4. Story bank — 6-8 STAR-format stories covering leadership, conflict, failure, scaling, cross-functional, and influence without authority. Pull from real career events, not generic templates. 5. Smart questions to ask — 8-10 questions sorted by audience (hiring manager, peer, recruiter, VP). No "what's the culture like" filler. 6. Honest read — where will they push hardest, what's the risk in my profile they'll probe, how should I get ahead of it. Voice rules: direct, no em-dashes, no AI tells, no buzzwords. Write the way I write.
Tools
Claude Projects Web search Conversation search
Impact

I walk in with the specific phrasing of the cultural questions, the right stories cued up, and a clear point of view on the role. No improvising on the things that matter.

03

Resume tailoring Claude Project

The problem

Every application needs a different resume. Doing it by hand for 20+ roles is unsustainable. Doing it with generic AI produces output that reads like AI and gets filtered out.

What I built

A persistent Claude Project that acts as my career coach. I paste a JD or share a URL and it produces a tailored resume in my voice with the original formatting preserved, saves the .docx to my Drive folder with a consistent naming convention, and updates my Excel application tracker with the role, fit score, status, and date.

Rules built in

No em-dashes. No fabricated achievements. No inflated titles. Match the structure of my base resume exactly. Bullets stay punchy and metrics-first.

See the prompt
You're my career coach and resume strategist for this job search. When I share a JD (pasted or URL), do this: Step 1 — Read the JD and identify the top 5-7 skills, themes, and keywords the employer is actually prioritizing. Rank them. Step 2 — Score my fit 0-100% against my base resume. Flag what's a strong match and what's a stretch. Step 3 — Tailor my resume: • Match the structure of my base resume exactly (same headers, same order, same formatting) • Rewrite bullets to align with the JD's priorities, only using achievements I actually have • Lead with metrics where I have them • Keep the voice mine: casual but direct Step 4 — Hard rules: • No em-dashes anywhere • No fabricated achievements or invented metrics • No inflated titles • No AI tells (no "leveraged," "spearheaded," "passionate about," "synergies," "robust") • No filler adjectives Step 5 — Save the output as a .docx to my Drive folder "Job Applications 2026" using: Luciano_Stephanie_[Company]_[Role].docx Step 6 — Update my Excel tracker: company, role, fit score, date submitted, status = "Applied".
Tools
Claude Projects Google Drive connector Excel tracker
Impact

A tailored resume in under 10 minutes per application instead of an hour. Output reads as written by me, not by an LLM.

04

Job matching, fit scoring

The problem

Not every role is worth applying to, and the ones that look like a stretch on paper sometimes aren't (and vice versa). I needed a faster way to decide.

What I built

The analytical layer that runs before the tailoring. It reads a JD, identifies the top 5-7 skills and themes the employer is actually prioritizing, scores my fit 0-100%, and gives me a brief honest assessment that flags whether the role is a reach and why.

Example

Scored 82% for a Senior Channel Partner Manager role at a SaaS partnerships platform and flagged the agency-side experience overlap as the strongest match, even though affiliate marketing was outside my HR AdTech background.

See the prompt
Read this JD and tell me honestly whether it's worth my time to apply. Output format: 1. Fit score (0-100%) with a one-sentence rationale. 2. Top 5-7 skills, themes, and keywords the employer is actually prioritizing, ranked. 3. My strongest matches against those priorities. Cite specific roles or achievements from my career. 4. Where I'm a stretch and why. Be specific about whether the stretch is closeable in a cover letter or whether it's a real gap. 5. Final recommendation — apply / pass / apply only if [condition]. Don't soften the read. I'd rather pass on a bad fit than waste a tailoring cycle.
Tools
Claude Base resume in context Career history loaded
Impact

I only spend tailoring time on roles where the fit score and the honest read both make sense. Cuts wasted effort and helps me prioritize the pipeline.

05

A custom Claude skill

Codifying my partnership outreach and account management playbook into a reusable Anthropic skill, so any new Claude conversation can pick up where I left off without me re-explaining the context every time.

06

A Claude Code workflow

Moving my Redshift SQL analysis routines into Claude Code so I can run pipeline performance reads from the terminal on demand, instead of building one-off Excel exports every time a question comes up.

07

A daily HR AdTech competitive intelligence brief

Same pattern as the Gmail brief, but pointed at the industry. A scheduled morning read that scans recruiting newsletters, funding signals, and key player moves to flag what shifted in HR AdTech yesterday.

The kit I keep coming back to.
Claude
Desktop, Projects, Code, and Cowork. The main tool for everything I build.
Granola
AI meeting notes. Replaces the need to take notes in real time.
Whisper
Voice-to-text. Useful when I'm thinking out loud away from a keyboard.
Notion
Where everything that isn't a Drive doc lives. Trackers, lists, draft writing.
Google Workspace
Drive, Gmail, Sheets. The layer underneath everything else.
AI &
building
  • AI Snack Club — fast, accessible AI updates I can read between meetings.
  • One Useful Thing — Ethan Mollick's essays on working with AI. The closest thing to required reading.
  • Every — Dan Shipper and team on AI in the wild, with depth.
  • Latent Space — when I want to go deeper on the technical side.
  • Anthropic Cookbook — official examples. Where I go to learn what's possible.
Recruiting
& HR Tech
  • Recruiting Brainfood — Hung Lee's weekly. The canonical newsletter for anyone in TA tech.
  • Chad & Cheese — long-running podcast on HR tech industry news and gossip.
  • HR Brew — a daily catch-all for what's moving in HR.