- 🔭 i build evaluations and tools for understanding when ai systems can be trusted. plicara labs is my independent ai research practice: reproducible studies, benchmarks, and practical tools.
- 💼 lead ii data scientist at s&p global, working on document extraction and internal benchmarking for model evaluation. previously, i worked on experimentation and customer segmentation at zillow and held data-science leadership roles at coca-cola femsa.
- 📄 applied mathematics at itam; msc in data science coursework completed in 2022, with my bayesian-optimization thesis expected in september 2026.
- 🌱 learning econometric theory, systems design, and rust (notes here as i go).
personal website · experience / cv · research
plicara labs — my independent ai research lab. how do we know an ai system works? i investigate that question through reproducible studies, benchmarks, and practical tools.
|
evaluate a regex the way a benchmark should — dfa equivalence with a shortest witness, correctness, and redos safety, because passing the tests is not the same as being right or being safe |
🔁 labloop
keep a change only if it measurably helps — an experiment loop for agent-driven research that runs trials under a wall-clock budget and records the reverted attempts too |
|
my study of generated regexes: test passing, vulnerability screening, and an audit of benchmark answer keys. read the study and its limitations. |
✏️ articles
notes and write-ups from the lab — recent research lives here and on plicara.ai |
- 🎯 bayesianoptim_bass — my msc thesis on spline surrogates for bayesian optimization, with companion research into how experimental machinery affects optimizer comparisons.
- 🧬 mollify — a rust-native static analysis engine for python that i help build with favio vázquez. reproducible findings for developers, ci, and coding agents, with explicit scope and confidence levels.
- 🤖 agentic engineering — a library of harness-agnostic agent skills: a cli-first data science pipeline, written once and distributed to any agent that supports the standard
- ⚡ loadstar — a go-based api for speed testing websites
- 🦀 rust learning — notes and code as i learn rust
- 🌐 adriantj.github.io — my personal site, blog, and project archive
languages
ai, ml and data
happy to collaborate on agentic systems, benchmarking, ai research, and anomaly detection.
