About
I'm an indie researcher and developer working at the intersection of language tooling, bioinformatics, and the kind of small-scale software that solves specific problems for specific people.
What I'm working on
LinguistPro is an offline-first Hebrew learning workspace built around a single conviction: a learner's data should stay with the learner. The architecture is unusual — SQLite WASM running in browser-local OPFS — and it lets the app deliver server-scale functionality without ever holding user data on a server. It's preparing for ulpan-cohort testing as part of a Hebrew-pedagogy diploma project.
EgorGenom is whole-genome sequencing analysis for my son, run in parallel with his clinical team. I built it on top of nf-core/raredisease and have been gradually adding domain-specific layers — phenotype-to-gene scoring, transcript-aware variant prioritization, live-database verification. The project is methodologically strict because the cost of being wrong is real.
Smaller projects and experiments live alongside these in my projects folder. Most of them are tools I needed to build because nothing existing did the job well enough.
How I work
Thorough over quick. I prefer enriching, complete paths over subjective top-N shortcuts. I distrust local databases when live ones are available, prefer reproducibility over cleverness, and update documentation in the same session as the work it describes.
The stack I default to: Linux/WSL, Docker, Nextflow for bioinformatics, Node.js + vanilla JS for web (LinguistPro is intentionally framework-free for debuggability), Python for analysis scripts, Astro for static content like this hub.