Faoot’s Lab
  • Jean Carlos Faoot Maia
Categories
All (4)
bayesian (1)
data-science (1)
inference (1)
llm (2)
machine-learning (2)
mcglm (1)
microeconomics (1)
nvidia (1)
optimization (1)
pricing (1)
python (1)
statistics (2)

Rethinking Generalization Error: From Double Descent to Test-Time Compute

statistics
machine-learning
llm

From the classical bias–variance tradeoff to overparameterized models, delayed generalization, and the compute budgets of modern language models.

Nov 17, 2026

Prefill Phase

Breaking the Decode Wall: Maximizing Local NVIDIA GPUs with Hermes-XLR

llm
inference
machine-learning
nvidia
optimization

Prefill, decode, and the memory wall — an optimization-first runtime for local LLM inference on consumer NVIDIA GPUs.

Jul 2, 2026

Price, Curves, and Agents

microeconomics
pricing
bayesian
data-science

Demand curves, elasticity, causal inference, and agentic pricing loops: the microeconomic foundations behind modern pricing systems.

Apr 3, 2026

MCGLM, a unified model for statistical analysis.

statistics
mcglm
python

An introduction to the Multivariate Covariance Generalized Linear Model (MCGLM) and its Python library for fitting multivariate and non-independent responses.

Feb 6, 2023
No matching items