Nestor Salamon.

I make better questions cheaper to ask.

I turn hard-to-operate data — images, signals, numbers, loose text, legacy systems — into tools people can actually explore. Whoever’s starting out reaches what they couldn’t before; the expert sees what always took too long to find.

Now Database systems & optimization Lowering the price of the question Data & AI at Appus HR
Portrait of Nestor Ziliotto Salamon

Report-first

“Which report do you want?”

Only returns variations of what already exists.

Decision-first

“What do you need to decide but can’t see?”

Opens questions nobody had a way to ask before.

01 The idea behind the work

When trying gets cheap, people try the impossible.

At TU Delft, every edit to an image or video cost hours of manual work — and when trying costs hours, nobody tries twenty times. So I turned those experiments into simple, real-time controls on ordinary hardware.

Data is the same: the information rarely arrives in a table and days of analysis is too expensive - some questions are never asked. Sometimes seeing which question is stuck is the whole fix, no tool needed. When one is, I build it.

The tool decides what gets perceived.

Change the tool and the same data says something else.

02 Different formats of the same problem

Sound, image, games,
databases, LLMs.

Chasing data that doesn’t fit in a table took me through every format it hides in. Different domains, one instinct: make the invisible perceptible, and the expensive cheap.

  1. 2020 — Now

    Data & AI Lead

    Appus HR

    Modeling and optimizing databases, pipelines, visualization and context engineering for people analytics.

    PythonPostgreSQLMySQLMachine LearningDatavizLLMsLooker

  2. 2024 — 2026

    Adjunct Professor

    Atitus Education

    Taught databases, database optimization and computer graphics. Also designed the course plans.

    PythonOracleSQLiteOpenCVOpenGL

  3. 2016 — 2020

    PhD Candidate

    TU Delft

    Real-time optimization algorithms for editing image and video — computational photography.

    MATLABC++OpenGLOpenCV

  4. 2015 — 2016

    Sr. Programmer Analyst

    Pmweb

    Web services and data integration for a hotel booking engine and PMSs.

    C#MVCSQL ServerETLDataviz

  5. 2010 — 2015

    Software Designer

    Hewlett-Packard R&D

    Web ERP and business intelligence for R&D finance and management operations.

    C#ASP.NETSQL ServerBusiness Intelligence

  6. 2008 — 2010

    Undergraduate Researcher

    PUCRS

    Web standards and Web 2.0 — courses and talks on emerging tech.

    HTMLCSSJavaScriptAjax

03 Education

The technical foundation
behind the tools.

TU Delft

2021 · TU Delft

PhD in Computer Graphics

Post-production image and video editing, realistic content generation and virtual exposure.

Image processingC++OpenGL

PUCRS

2015 · PUCRS

Master in Computer Science

People and group re-identification using color descriptors and soft-biometric attributes.

Computer visionC++OpenCV

PUCRS

2012 · PUCRS

Bachelor in Computer Science

Multimedia sound generation based on Newton’s method for equation solving.

JavaMIDIOctave

04 Making expensive things cheap

Selected
publications.

Exposure, motion and light — each paper turned something that cost hours of manual work into a real-time control anyone could use. All publications & projects

Computer Graphics Forum · Pacific Graphics · 2019

ShutterApp: Spatio-temporal Exposure Control for Videos

N. Z. Salamon, M. Billeter & E. Eisemann

Paper

TVCG · 2020

Controllable Motion-Blur Effects in Still Images

X. Luo, N. Z. Salamon & E. Eisemann

Paper

Computer Graphics Forum · Eurographics · 2017

Computational Light Painting Using a Virtual Exposure

N. Z. Salamon, M. Lancelle & E. Eisemann

Paper
05

Life, not a footnote

There is always
another angle.

Same track, different headphones. Same place, different lens.

A personal moment
A personal moment
A personal moment
A personal moment
A personal moment
A personal moment
A personal moment
A personal moment
06

Say hello

Stuck by a tool
nobody built yet?

If your team stops at the easy questions because the good ones cost too much, we probably have something to talk about.

BibTeX

Cite this paper