The dailies
Daily log
What each day actually held, one line at a time. Names are replaced with roles and internal details are left out; the private notebook keeps the rest.
- First day. Team standup, met my manager and my co-intern, worked through onboarding trainings.
- Read the project scope and booked my first meetings.
- Onboarding buddy intro, new-hire sessions, intern welcome lunch, badge pickup.
- First questions to my mentor about the project spec and the AI tooling.
- Culture 101 and intern cohort sessions; met two more engineers on the team.
- Started reading the platform's authoring and replication docs.
- IT onboarding sessions, plus a readout directly relevant to how gated content flows through my project.
- Cloned the repository, set up the build workspace, started the local environment.
- First manager one-to-one.
- Local setup got tangled badly enough that restarting from scratch was faster. Logged it as a loss.
- Worked through the replication API and the content model.
- Set the week's goals: local environment plus authoring-UI familiarity.
- Configured local development credentials, then walked the authoring UI the way an author would.
- Mentorship meetings and intern one-to-ones; met more of the team.
- Local environment fully running. Installed the CLI tooling, set up sprint tracking, incorporated reviewer feedback on the project plan.
- The day the project changed shape: the feature already existed, so the work became extend, not build.
- Wrote the design document end to end, built three architecture diagrams, chose the technical approach.
- Manager one-to-one and a learning session.
- Design-doc polish; built the risk register and the open-questions list.
- Independent refinement day.
- Documentation pass on the codebase notes; mockup iteration on badge placement inside the 72px bar.
- Documented the full codebase map and ran the research phase as five parallel investigations, each verified against source.
- Fixed the diagram overlaps. Scoped the hardest signal down to a first-release version and recorded why.
- AI program kickoff and a spec review with the PM.
- Built this tracker: the dashboard, the daily log, and the journal CLI.
- Recovered the research doc after a tooling failure ate it.
- Design review with my mentor surfaced five discussion points.
- Sat in on an operational postmortem readout, useful for seeing how the team separates root cause from symptom.
- Folded the design-review feedback into the doc and reworked the mockups.
- Documented the gating signals and the wording constraint that follows from them.
- Holiday, Independence Day observed. No work.
- Sprint planning and team demos.
- Resolved all five design-review points into the doc, each verified against the codebase.
- Week-4 deliverable hit: design document complete and reviewed with the team.
- Set up the design-tool integration and iterated on the mockup.
- Prepped the design-review invite with context so nobody walked in cold.
- Built the HTML and CSS mockup from scratch: the editor shell, five status badges, and the acknowledgment dialog with a publish-drift timeline.
- Recreated it in the design tool for side-by-side comparison.
- Week-4 deliverable hit: dialog wireframes done in both tools.
- Weekly manager one-to-one, and an intern cohort sync I organized.
- Mockup refinements: badges stacked on the right edge, deactivated state added.
- Broke the project into tracked tasks with copy-ready descriptions.
- Built a humanizer pass and cleaned the whole journal with it.
- Ran the team design review. 38 comment threads triaged and all 29 agreed edits applied the same day.
- Traced the content-gating stack end to end. Mockup revision 3: click-only panel, production fidelity.
- Mentorship check-in and a high-level design review of the backend approach.
- Sketched the endpoint response contract for the first change.
- Consistency review across all three project docs; applied all 14 recommended edits.
- Refreshed the tracker to the revised requirements. Exam practice at the AI program office hours.
- Organized and ran design review 2.
- Researched endpoint registration options ahead of implementation. Added the project to my resume.
- Weekly one-to-one.
- Drafted the first change's approach on paper and wrote out its failure modes before writing any code.
- Sprint planning.
- Read through the codebase to plan where the changes land, and wrote up where each requirement lands in the code.
- Mentorship check-in and review scoping against the design doc.
- Mockup pass on the two edge-case states.
- Out sick most of the week. No implementation progress, logged as a loss.
- A light documentation pass was all the energy allowed.
- Broke the second half of the project into tracked subtasks.
- Midpoint review prep: pulled the story together from the docs and mockups.
- Completed the midpoint self-review form.
- Built the midpoint review deck from the mockups and the design doc.
- Planned the implementation phase as four sequential code reviews, each standing on its own.
- Midpoint review with my manager: walked the self-review and the project progress.
- Started the first change on a branch and drafted the test matrix before the tests.
- Mentorship meeting.
- Wrote the sync-status and attribution logic by hand; drafted test cases against the endpoint contract.
- Rebuilt the task board: nine scoped tasks with copy-ready descriptions.
- Self-review pass: read my own diff looking for my own bugs before anyone else saw it.
- Weekly one-to-one, then turned the plan into a readable task board.
- Backend unit tests written out.
- Got the full project running locally, every module active.
- Verified the endpoint against the running instance rather than only against mocks.
- Sprint planning. The first change committed on its branch.
- Merged both review passes into one fix doc with no duplicate findings, and migrated the project docs to the team's newer doc tool.
- First-reviewer findings verified with the suite green.
- Mentorship meeting, and a leadership ask-me-anything for the intern cohort.
- Applied every second-change review finding and re-ran both test suites.
- Weekly intern cohort meeting.
- Drafted the second change's fixes by hand from the review findings.
- Adversarial review on the second change: eight verified findings, one a genuine fail-open defect.
- Wrote the review doc so the second reviewer starts warm, including the eleven refuted candidates.
- Weekly one-to-one.
- Implemented the fixes, cleared the leftover TODOs, and added a solutions doc pairing each bug with its fix.
- The second change committed: the badge layer on the editor bar.
- Completed the final self-review and prepped the final presentation narrative.
- Tallied my AI-assistant usage for the summer from local session data.
The ledger
Meetings and one-to-ones
Every meeting logged, by role and date. Who they were with stays in the private notebook.
The brief
Project brief
n editor bar for AWS Marketing's publishing platform that keeps authors honest about publish state: whether the page matches production, whether someone else changed it, and whether it is live yet.
Goals
- A visual indicator that stays in sync with the live website
- Surface whether someone else made changes
- Show whether the page is live yet
Stack
Deliverables
- Bar states: in sync (page matches what is published), out of sync (modifications since last publish), and scheduled (launch-sensitive content that should not be published yet)
- Last modified: who and when
- A notification flow when you open a page someone else changed since last publish, with an explicit Acknowledge and Continue step
Planning
Weekly goals
The plan the internship ran on, week by week.