I'm Anya Lazarenko. I spent 9 years at Snap leading technical programs across product, engineering, and operations. Today I also design and build AI products myself — two of them are live below, and you can open both right now.
I designed and built both of these products myself. They are live now — open them and try them.
Scan to try it on your phone
Visible Local · Live
AI visibility for local businesses
Visible Local checks whether AI search and chat assistants mention a small local business when people ask for what it offers. When the business is missing from the answers, it shows why, and what to fix, in plain language.
LexiQ turns anything you want to learn into flashcards: your own decks, CSV import with an AI prompt template, pronunciation, and SM-2 spaced repetition. No account, no backend — free forever, installable on any phone.
Most AI products are built for people who already work in tech. I build for everyone else: the small local business that needs customers to find it, and the person learning a new language for a new job or a new country. When AI search recommends a neighborhood business, that business can reach customers it never could before, and when a learning tool is free, anyone can start. I want the skills I built in big tech to do that kind of work.
9 years
at Snap Inc., leading technical programs
PMP · CSM
Project Management Professional & Certified ScrumMaster
2 master's degrees
Economics and Computer Science
Website Director
PMI Los Angeles chapter
04 / More experiments
More experiments.
Open-source projects and studies, with the code on GitHub.
Legal Answer Evaluation · Shipped
How good are AI legal answers, really?
A reproducible evaluation study: 40-item benchmark with gold references, anchored rubric, independent model judges across vendors plus a human rating pass, and inter-rater agreement statistics.
40 gold-reference items, an anchored rubric, independent judges across vendors plus a human rating pass — everything needed to rerun the study from scratch.
Four Claude Code agents support work breakdown and dependencies, risk scoring, weekly status reporting, and release go/no-go recommendations. The PM decides. The agent proposes.