Chinh (lelouvincx) / Weekly Report - 2026-W11

Created Mon, 25 May 2026 00:00:00 +0000 Modified Mon, 25 May 2026 06:02:25 +0000
1532 Words
page

Weekly Report — Mar 9 – Mar 13, 2026 (friday mode)

Pull Request: #7 and #8

  • Work

    Internal

  • (done) Code review MRR PR #812 — completed 4 review rounds on customer identity resolution refactor (DAT-471). Decomposition of itg_mappings__customer_identities into 5 sub-models. Identified critical issues: Zoho-only rows silently dropped by WHERE clause, NULL domain NOT IN semantics, Zendesk fan-out risk. PR merged Mar 10. (PR #812)

  • (done) Continue commenting DE-201 (PR Workflow enhancement) — completed review and documentation of CI/CD gaps in internal-aml-project repo. Proposed tiered roadmap: quick wins (PR templates, CODEOWNERS), expanded test coverage, new automation (lineage diff, impact analysis). (Notion)

  • (done) Adhoc fix failed CI upload-holistics in dbt project — Ruby 2.7.4 EOL broke activesupport gem install. Migrated to new standalone Holistics CLI binary; modernized workflow with uv for Python/dbt. (PR #838, merged Mar 9)

  • (done) Review PR #839: CI dataset scope — changed CI dataset creation from commit-scoped to PR-scoped for better debuggability. Found critical shell syntax error, uniqueness loss in fallback, and concurrency risks. All resolved. (PR #839, merged Mar 9)

  • (done) Check Prefect Slack automation failure — Dependabot upgrade broke a package used by Prefect for Slack notifications. Simplified pyproject.toml, debugged flows, identified root cause. (PR #377, merged Mar 12)

  • (done) Fix DAT-554 Elementary dbt test n_records column missing — fixed compatibility issue in custom unique test for Elementary. (PR #841, merged Mar 13)

  • (done) Check if current API supports the Pass The Keys ticket — investigated feasibility of supporting object dependency data sharing via Holistics API. Continued from duty support ticket.

  • (in-progress) Plan for Census migration to Fivetran — documented migration plan for reverse-ETL jobs (BigQuery → HubSpot). Census sunset Apr 1; confirmed MAR usage (~1,215 records/month) fits Fivetran Free plan (3,500 MAR). 2 syncs to validate: dim_contacts → HubSpot Contact, fct_customer_stats → HubSpot Company. Next: create Fivetran account under [email protected], migrate, validate. (Notion)

  • (in-progress) Check failed dbt tests in #data-ops-bot — listed all issues (DAT-554, DAT-555). Analyzed fct_job_queue_performance model: stale, superseded by squad performance dashboards; recommended removal. mart_product__dataset_datamodel_dimensions also flagged. Next: create task/issue for each observation.

  • (in-progress) DE-202 trial-first vs call-first conversion — received assessment from growth team: several customer categorizations incorrect (Fanserv, Novigi are trial-first not call-first; Basata reassigned to Nam). 11 customers identified as not trial-only. BizOps team has answered; next step is to review and plan tasks. (Notion)

  • (done) DE-204 Auto-enable AI for trial accounts — updated field references (public.tenants.settings->enablement, public.users.settings->personal->ai->hide_banner). Found existing AI Settings dashboard. Done: confirm if adding hide_banner tracking meets requirements. Next: no action.

  • (in-progress) Listing and re-prioritizing pending Linear issues — pulled 15 open issues (DAT-450, DAT-469, DAT-510, DAT-524, DAT-155, DAT-163, DAT-219, DAT-259, DAT-268, DAT-283, DAT-371, DAT-462, DAT-463, DAT-504, DOC-36). Plan to triage and complete small ones.

    Duty Support

  • (done) Data support for Trust Wallet — customer needed help with AQL metrics for weekly unique users on swap data. Impersonated into account to investigate. (Slack)

  • (done) Support Pass The Keys sharing object dependency data — created shareable dashboard, suggested dataset custom views. Customer replied that migrating to 4.0 is cost-prohibitive. (Zendesk #24674)

    Presales

  • (done) Add tabs to Hooli dashboard — solution proposal completed. (Slack)

  • (done) Plan building a live demo dashboard for overview slide — prototyped in Gemini, implemented via Ampcode. (Slack)

  • (done) Feedback — requested UI for Query Model Persistence (Aarki customer). (ReadAI)

    Teaching

  • (in-progress) Complete Brazilian dashboard on Metabase for Yen — Dash 1 (Sales Overview) done. Dash 2 (Shipping Overview) and demo video recording still TODO.

  • Personal / Tooling

  • (done) Built SmartClass website for thay Thanh — static site with shadcn/ui, deployed to Cloudflare Pages (smartclass-demo.pages.dev). Constraints: deployable on Cloudflare with minimum cost, as simple as possible. (PR #1, merged Mar 8)

  • (done) SmartClass v0.1 — project setup with wrangler.toml, D1/R2 bindings, Hono worker entry. D1 schema migrations (users, exercises, answer_schemas, submissions, lectures). Auth: phone+password login/register, JWT middleware. React Router, login/register pages. Deployed via Cloudflare Pages + Workers.

  • (done) Fixed Logseq weekly report automation timing — runs were triggering at wrong times.

  • (done) Added Linear CLI skill for Amp — using schpet/linear-cli.

  • (done) Learned how to query Logseq blocks — figured out pattern ((uuid)), just grep it, no complex tool needed.

  • (done) Clean up Arc tabs.

  • Noted: tooling investment paying off (4 agent skills set up last week). Should consider cooling-off period before new skill builds.

  • IDEA: Do self “performance review” every 3 months. Next target: end of March.

  • IDEA: Archive all Amp threads into a repository — similar to claude-mem. Motivation: backup against context loss.

  • Explored opencode with GPT-5.3 Codex model.

  • Learning & Notes

  • Memory Decreasing — self-reflection on declining memory. Likely causes: sleeping only 6 hours/day, excessive social media scrolling reducing attention span. Plan: reduce social media, set phone to grayscale, take breaks every 1–2 hours. Related to recurring “Down The Rabbit Hole” pattern.

  • Dominant Eye — discovered right-eye dominance affects screen layout preferences. Explains why dock on left felt uncomfortable. Set up monitor arrangement accordingly: primary monitor centered (right eye anchor), secondary at 30° left angle.

  • ThePrimeagen’s “99” (GitHub) — a tool bridging hand programming and full vibe coding. Pitch: people want AI help but not 100% agent control (grep/edit all files). Interesting middle ground.

  • BI Chatbot exploration — researched how a normal data team would adopt a BI chatbot. Candidate: nao.

  • Random thought: should follow the AI trend rather than building something from the last era.

  • Next Week

    From pages/Backlog.md, ordered by priority:

    P1

  • Presales: Onboarding call with Showbie

  • Presales: Onboarding call 3 with Basata (Slack)

    P2 — Should Do

  • Internal: Fix excluding internal testing Zoho accounts (DAT-524) — lingering since week 7

  • Internal: Continue Cross-Model Calculation — timing may still be premature (docs revamping not far enough)

  • Internal: Upgrade dbt-core of repo dbt to same version as Prefect

  • Internal: Prepare for phase 2 MRR problem

  • Internal: DE-202 — review BizOps team’s answers and plan tasks (carry-over)

  • Internal: DE-204 — confirm hide_banner tracking meets requirements (carry-over)

  • Internal: Check failed dbt tests — write Notion doc for fct_job_queue_performance removal, let teammate review (carry-over)

  • Add2Cart: Document Add2Cart data model; project retro with Dong

  • Duty Support: Document Medenterprises performance use case — overdue since Feb 5

  • Teaching: Figure out set-based thinking curriculum; find football dataset for Duc Anh; checkout pudding.cool; research visualization resources

  • Teaching: Complete Brazilian dashboard Dash 2 + demo video for Yen (carry-over)

  • Personal: Consider a Flashcard app for learning English words

    P3 — Nice to Have

  • Presales: Read Modeling Patterns docs; watch Taxdome onboarding call; learn embedded analytics use case; compare ChatGPT solutions Omni vs Holistics

  • Internal: Triage and complete small Linear issues (15 open)

  • Duty Support: Check with Product on metric hover detail; document Tricia’s A/B comparison use case

  • Personal: Search for YouTube video summarization tool; self-host readeck; migrate remaining Obsidian notes; edit video for 4/11

  • Misc: Think about how a normal data team would adopt a BI chatbot

    Carry-over from this week (in-progress)

  • Census migration to Fivetran — create account, migrate, validate (deadline: end of March)

  • DE-202 trial-first vs call-first — review BizOps answers, plan next tasks

  • DE-204 AI enablement tracking — confirm requirements with chi Phuong

  • Brazilian dashboard for Yen — Dash 2 + demo video

  • Career & Personal Consulting

    Observations from this week:

    1. Deep infrastructure work: This week was heavily infrastructure-focused — 4 dbt PRs (merged), 1 Prefect PR, CI fixes, Elementary compatibility, Census migration planning. This is unglamorous but high-leverage: each fix prevents cascading failures across the data platform. The DAT-554/555 investigation showed good judgment — instead of just fixing the test, you analyzed whether the underlying model (fct_job_queue_performance) was still needed and recommended removal. This “fix or kill” mindset reduces maintenance burden.

    2. Linear backlog triage is overdue: You pulled 15 open Linear issues and acknowledged they need prioritization. Several (DAT-524, DAT-450, DAT-259) have been open for months. The intent to “complete them as soon as possible” is good, but be realistic — batch the truly small ones (DAT-219, DAT-268) into a single focused session, and explicitly defer or close ones that are no longer relevant. Carrying 15 open issues creates cognitive load.

    3. Memory and attention concern: The “Memory Decreasing” reflection is worth taking seriously. Sleep deprivation (6 hrs/day for over a week) directly impacts working memory and decision quality. The plan to reduce social media and take breaks is sound, but the root fix is sleep. Consider a hard cutoff time and protect 7+ hours. This is a performance optimization with the highest ROI.

    4. SmartClass side project: Building a full-stack app (Hono + D1 + React Router + auth) in a weekend demonstrates strong execution speed. The “vibe coding” approach with constraints (Cloudflare, minimal cost) was pragmatic. Watch that this doesn’t become another rabbit hole — set a clear scope boundary for v0.1 and park it.

    5. Census migration deadline: Apr 1 is approaching. You’ve documented the plan and confirmed pricing fits — now execute. This is a clear P1 that should be completed next week. Don’t let it slip to the last minute.

    6. Breadth continues: Internal (7 items), Duty Support (2), Presales (3), Teaching (1), Personal (6). The pattern of touching many projects persists. This week the infrastructure work was necessary and well-executed, but the number of context switches is high. Consider blocking 2–3 “focus days” next week for the Census migration and Linear backlog triage.