Statistician, entrepreneur and AI researcher.
I have been building things for over 20 years. Today I lead Synerise in Brazil, do an MSc in AI at UNICAMP, coordinate the Mensa Brasil AI special interest group and develop Gnostikon — a serious attempt to make the system not lie about what it knows.
Five hypotheses repeatedly confirmed.
I do not treat these as beliefs. They are positions that survived confrontation with real situations — small company, large company, failure, success, sale, founding, classroom.
- 01
Constructive non-conformism
The status quo is a hypothesis, not a fact. Mental laziness is the main adversary.
- 02
Radical autodidactism
Learning is not an activity — it is the infrastructure.
- 03
Truth & rationality
Epistemic integrity over appearance. I dislike what tries to seem more than it is.
- 04
Happiness through the problem
Satisfaction is not in the answer — it's in formulating the question well.
- 05
Exemplary ownership
If a teammate fails, the failure is the system's. Treat others as you would be treated.
From UNICAMP to Synerise — in reverse order.
- 2025–
Country Manager Brazil
· SyneriseLeading the expansion of the Polish behavioral-AI company in the Brazilian market.
- 2024–
Founder
· Heitmann AIOwn research and products around AI agents, alignment, and knowledge canonization (Gnostikon).
- 2024–
AI SIG Coordinator · Brazil
· Mensa BrasilCoordinator of the AI special interest group, with monthly sessions and open inquiry.
- 2022–2024
Director of Innovation
· Quality DigitalAfter ACCT's sale, led innovation and international expansion for the group.
- 2009–2022
CEO & Founder
· ACCT Consulting & TechnologyGrew a small Atibaia team into a 140+ person consultancy serving Unilever, AbInBev, Carrefour and Reckitt across 25 countries. Sold to Quality Group in 2022.
- 2018–2023
Faculty
· EICOM Institute · Cambridge / BostonImmersive executive program on digital transformation.
- 2008–2009
Risk analyst
· Itaú-UnibancoRisk modeling for the pension portfolio.
- 2007–2008
Planning analyst
· ElektroTime-series forecasting for energy demand.
- 2003–2007
BSc Statistics
· UNICAMPWhere it all started: estimators, maximum likelihood, and the first intuition that data does not speak for itself.
What's on the table today.
- LOCATION
- São Paulo · Brasil
- ROLE
- Country Manager Brazil @ Synerise
- BUILDING
- Heitmann AI · Gnostikon
- STUDYING
- MSc in AI · UNICAMP
- READING
- Hofstadter — I Am a Strange Loop
- NEXT
- VTEX Day 2026 · Apr 16-17 · panel