Data leader across analytics and data science
I scaled and led an 8-person analytics and data science team in Brazil.
I owned a data platform end to end, from the warehouse and dbt models to the BI layer.
Career history on LinkedInTry it: hover or tap a point
How many accounts a trial needs
as users inside one account behave more alike
Each study leads with its question and its answer. Open one for the method, the evidence and the limits behind it.
Planning, on hypothetical assumptions: test activation first, sized for its own rollout rule
How should successful tasks connect to a retention hypothesis?
Test activation first. A trial sized for its rollout rule needs about 360 accounts.
Hypothetical planning assumptions; no measured business improvement.
HTML, Python, experiment planning
Open Enterprise AI adoption33of 40
correct SQL answers with full documentation; 27 from the bare schema, 31 with allowed values
How much does documenting data marts help an AI model answer questions in SQL?
Documentation raised correct SQL answers from 27 to 33 of 40. Listing the allowed values alone reached 31.
One model run on 40 related, purposively written questions over synthetic data, so the intervals describe this question set and support no general claim about documentation.
Python, SQL, DuckDB, LLM evaluation
Open Agent-ready data martsTo get in touch, write to hello@diogotravassos.dev or visit Diogo on LinkedIn.