Case Study
AI Transformation at Scale
Robinhood's lifecycle marketing team was using AI sporadically: a few people, a few prompts, no system. I built the system: an autonomous Claude Code agent that runs lifecycle campaigns end-to-end, a library of 40 reusable skills and sub-agents, and a 2-day bootcamp that took the team from 20% to 100% weekly AI usage. Then I built the measurement framework to prove it, tracking real workflow behavior instead of training completion.
- Autonomous agent running lifecycle marketing campaigns end-to-end, cutting multi-week workflows to hours.
- 40 reusable skills and sub-agents spanning planning, targeting, content, QA, and performance tracking.
- 2-day AI bootcamp, pitched and run end-to-end, taking weekly AI usage from 20% to 100%.
- Adoption measurement framework tracking workflow behavior against defined baselines.
- AI transformation roadmap sequencing rollout across the org by team readiness and tool maturity.
Claude Code · agentic sub-agents · MCP integrations · adoption analytics
Personal Project
Job Search, Automated
When I started my own search, I treated it like any other ops problem: too much manual, repetitive work standing between me and the part that actually matters. So I built three Claude Code skills to run it.
~/skills/linkedin-job-alert-digest
Turns a flooded inbox into a clean, deduped digest.
- Searches Gmail for LinkedIn job alert emails from the last 24 hours.
- Parses each listing and deduplicates by Job ID.
- Enriches new postings via browser automation, pulling full description and salary range.
- Cross-references against the job tracker to skip anything already seen.
- Appends new jobs to the tracker and outputs a digest grouped by alert.
$ /linkedin-job-alert-digest
~/skills/application-status-check
Keeps the pipeline honest without manual digging.
- Reads the active pipeline from the application tracker.
- Searches Gmail for each tracked company in the last 24 hours.
- Detects rejections, interview invites, recruiter screens, offers, and assessments.
- Updates status and appends a dated note, flagging anything ambiguous for manual review rather than guessing.
$ /application-status-check
~/skills/tailor-resume
Builds a tailored resume and cover letter for every role marked "Interested."
- Finds postings marked Interested that don't have materials yet.
- Pulls the live job description directly from the source.
- Rewrites the summary and reorders bullets to mirror the job description's language.
- Generates a tailored resume and cover letter as .docx files, and updates the tracker to Resume Ready.
$ /tailor-resume
Full source → github.com/jdizz7420/Job-Search-Assistant