Chinh (lelouvincx) lelouvincx

• Done

  • DONE Review test cases again (50%)

Note

• https://www.reddit.com/r/vibecoding/comments/1myakhd/how_we_vibe_code_at_a_faang/
  • Always start with a solid design document and architecture.
  • Build from there, always write tests first.
  • It’s not too different from how Holistics currently works.
  • Implies that the best way to “vibe code” is not to vibe code at all.
  • I slowly get out of the thinking that AI can plan and code 100% without human intervention.
    • Looking back at my thesis, which was 100% AI-generated, is totally trash.

• Done

  • DONE Write document for total_items in mart_product__tenant_features_usage_monthly

Done

• DONE Add dbt model for event tracking tagging system

• DONE Collect AI issue

Done

• DONE PR review https://github.com/holistics/prefect/pull/338

• Done

  • DONE Xin bản sao trích lục giấy khai sinh

• Note

  • Options for providing data to AI:
    • Metadata: titles, labels, descriptions, text blocks, data types, model relationships, formulas, chart settings, AMQL codes, git commit messages.
    • Sample data: Few sampled values of the source database columns.
    • Result data: result data (in form of csv or images) of the generated chart.

• Done

  • DONE Fix wrong tests

• Done

  • DONE Fix wrong AQL in tests

• Note

  • https://openai.com/index/introducing-swe-bench-verified/
  • Revise the goals of benchmarking:
    • P1 Internal: Track our AI’s improvement.
    • P1 External: Communicate (firstly with internal Holistics stakeholders: Sales team, Product team; then with prospects & customers).
  • This is divided into many tasks:
    • Creating test cases - I’m working on this.
    • Evaluation method.
    • Evaluation pipeline.
    • Presenting evaluation result.
  • Goal of creating test suite for benchmarking:
    • Make a gold standard for evaluating agents’ ability to work with data-related business problems, rather than generating SQL/AQL snippets (like text-to-sql models do).
    • Have a measurable way to indicate how good our AI is (in aspect of better support analytics techniques - running total, percent of total, trend analysis; reduce lines of code) that correlate with real data analyst productivity.
    • Ease of comparing Holistics’s AI with chatgpt, text-to-SQL models, or competitors (?).
    • Identify capability gaps (where all agents fail).
  • Expected one-line pitch:

    • Our latest AI agent can now successfully handle 95% of common analytical patterns like Period-over-Period comparisons and 85% of complex multi-step Cohort analyses, representing a 20% quarterly improvement in advanced analytical support, according to our benchmark dataset.

Done

• DONE Tìm bằng tốt nghiệp cấp 3

• DONE Test for window functions (50%)

• DONE Fill 10 AQL in existing tests

• DONE Review https://github.com/holistics/dbt/pull/794

• Note

  • eliot-jones.com
    • Beautiful because it’s simple, light, and focus on the content.
    • Love it.
    • Also https://bytemash.net/posts/i-went-down-the-linear-rabbit-hole