Academic services

I code qualitative data for research teams: interview transcripts, focus groups, open-ended survey responses. In ATLAS.ti, in Hebrew, English or Spanish.

Most of the work is coding interviews against a codebook, or building the codebook with you from the first interviews. I do this as a research assistant to Dr. Gal Levy at the Open University of Israel, on a funded project, and that is where the way of working described below took shape.

What I do

• Interview coding in ATLAS.ti. Against your existing codebook, or one we build together. You get the coded ATLAS.ti project and a case-status file: progress per interview, decisions pending, and the passages deliberately left uncoded, with their positions and the reason.
• Codebook development and maintenance. Definitions, scope rules, and a coding brief that records the decisions already made, written so that a second coder can be bound by it.
• AI-assisted coding with a human in the loop. The AI suggests from the live codebook, I decide, the AI writes only what was approved. Every code application comes with a definition-grounded argument, and the trail shows what was set aside and why. The method is public: the essay and the skill on GitHub. If your ethics approval excludes AI tools, I code by hand with the same documentation.
• Cross-language projects. Interviews in one language, analysis and writing in another. I work in Hebrew, English and Spanish, and I ran multilingual teams for twenty years.
• Workshops. A half-day session for a department or research group on AI-assisted coding in ATLAS.ti: what to automate, what to protect, and how to keep the methods section defensible.

How it works

1. A short call about the study, the codebook and the ethics constraints.
2. A paid pilot on one interview, delivered with its case-status file, so you can judge the coding before committing to more.
3. Then per interview or per project, at a price agreed in advance.

Data

Your data stays in your ATLAS.ti project. When AI assistance is used, the passages being coded are processed by the AI provider under its terms; we go through those terms together before the pilot, and AI is used only where your data agreement and ethics approval allow it. I never publish, quote or reuse participant data.

Background

Twenty years running customer support and operations teams, twelve of them at WPML. Research assistant on a funded sociology project at the Open University of Israel. M.A. in Interdisciplinary Democratic Studies in progress. More on the about page.

Contact

Write to amitkvint@gmail.com, or message me on LinkedIn. I answer in Hebrew, English or Spanish.