Applied AI / Machine Learning Engineer & Researcher
Rafael Luiz Testa, PhD
I develop and evaluate AI systems in computer vision, generative models, natural-language analytics, and machine learning engineering.
What I work on
I combine model development with data, software, and evaluation. I focus on failures that are easy to miss when a system is assessed only through its final output.
01
Generative Vision
Video synthesis, temporal coherence, facial analysis, and synthetic media.
02
Reliable AI Systems
LLMs, Text-to-SQL, evaluation, validation, and deterministic fallbacks.
03
Research Engineering
Reproducible experiments, data systems, backend engineering, and operational reliability.
Selected work
These projects cover generative video, facial analysis, deepfake detection, and LLM-based analytics. Each project began with a specific failure mode and required an implementation and evaluation designed around that failure.
Generative vision · Research
Temporal Coherence for Facial Video Synthesis
Treating temporal coherence as a first-class property in image-to-video facial-expression synthesis.
Geometry-aware vision · Research
Facial Expression Synthesis Based on Similar Faces
A geometry-aware computer-vision pipeline for retrieving a similar face and transferring an expression.
Synthetic media · Reproducible ML
Deepfake Detection Based on Ratio Images
Detecting manipulated videos through interpretable temporal differences between adjacent frames.
Applied AI · Research engineering
Reliable Natural-Language Analytics
A validation-first architecture for asking questions about public-sector analytical data in natural language.
Selected publications
My publications cover temporal coherence in generative video, synthetic-media analysis, facial-expression synthesis, and efficient facial-expression classification.
Current direction
In my current work, I develop an internal natural-language analytics platform for public-sector data. The system grounds questions in data semantics, validates candidate SQL, restricts database execution, and checks presentation choices independently.
Contact
I am available to discuss applied AI, research engineering, and technical collaboration.