Model Systems
Post-training harnesses, retrieval pipelines, telemetry budgets, and evaluation loops.
ML systems overviewZach Olivier
Machine learning systems, evaluation, and applied research.
I am a machine learning engineer working on production ML systems and research tooling: post-training, retrieval, anomaly detection, tabular modeling, reliability, and artificial life.
Applied Science
Reliability, survival analysis, anomaly detection, tabular foundation models, and uncertainty.
Research Tooling
Local-first paper ingestion, concept graphs, reproducible experiments, and benchmark reports.
Selected ML work
All projectsFine-tuning
Concensus SFT
NLP / supervised fine-tuning / evaluation Scientific QA SFT pipeline with data preparation, training artifacts, and loss diagnostics.Research systems
Second brain knowledge pipeline
LLM workflows / knowledge systems / agent review Local-first knowledge pipeline for papers, notes, reviewed article nodes, and concept graphs.Computer vision
Industrial anomaly detection
Anomaly detection / visual inspection / demo PatchCore-style industrial inspection lab with memory-bank retrieval and a live static demo.Applied modeling
Tabular foundation model lab
Tabular ML / foundation models / baselines TabPFN and TabICL lab against classical baselines with runtime, uncertainty, and benchmark reports.ALife