Identity
Andrew Zavala, PhD
andrew@gestalt-lab.com
Andrew’s work is rooted in a personal turning point that deepened his fascination with perception: the process by which the brain transforms raw input into intentional experience. He now brings that focus to AI vision, designing metrics and datasets grounded in psychophysics and human empirical data.
Daniel Lougen, M.S.
dan@gestalt-lab.com
Dan’s work sits at the intersection of visual neuroscience, model engineering, and agentic AI. Trained in perception science and working hands-on with fine-tuning, quantization, and multimodal systems, he builds tools and datasets that make models more robust, more interpretable, and more deployable in the world. Recent projects range from agentic curation pipelines and model resilience experiments to onboard vision systems for autonomous wildfire detection.
Follow Us on Hugging Face
Selected Systems
SYSTEM
CHIMERA - Agentic Data Curation with an 8-Factor Quality Model
Pipeline for ranking and filtering conversational training data using multi-factor quality scoring and gated reward aggregation.
SYSTEM
TALOS - Structured Redaction and Scoring for Agentic Data Pipelines
Pipeline for privacy-aware preprocessing, trace scoring, and dataset packaging prior to public or shared release.
SYSTEM
Ornstein - Robust Fine-Tuning and Quantized Deployment for Open Models
Experimental fine-tuning and quantization framework aimed at improving resilience while supporting efficient local inference.
Selected Publications
PUBLISHED
The Alpha-band of the EEG Modulates the Perceived Location of Moving Targets
Journal of Visual Science (2024)
PUBLISHED
Asymmetric Point Velocity and the Perception of Volumetric Depth.
Gestalt Intelligence Review (2024)
Works in Progress
IN PREP
Sampling Motion in Time: Does Alpha Entrainment Shift the Perceived Onset of Motion in the Fröhlich Effect?
Target: CVPR 2027
Perception meets practice.
