Liubov

Luba Kaper

Software engineer · New York

Download PDF

Summary

Software engineer with production experience at Twitter, now building AI-assisted software for small teams doing paperwork-heavy, regulated work. I work directly with the people doing the job, then design, build, test and deploy the system. I use AI where it saves time and keep records and decisions under clear rules and human review.

Experience

Software Engineer, Freelance

Aug 2026 – Present
NY Business Licensing · New York, NY · Google-funded Hispanic Federation program
  • Sole engineer on a Next.js, TypeScript and Supabase/PostgreSQL operations platform for a 4-person firm that gets restaurants through New York liquor license, food permit and sidewalk cafe approvals. It replaces spreadsheets, email, Dropbox and memory; the client projects it will save 5 to 10 hours a week once fully adopted.
  • Modeled the state liquor license process as an 8-stage workflow, from application sent to final license, with fixed rules for stage changes, task links and next steps. This part is plain code, not AI, so it behaves the same way every time.
  • Gave AI one narrow job: improving the wording of follow-up emails based on the application’s confirmed stage. It can’t change records, pick recipients or send anything. Staff review and send every message.
  • Ran discovery with the owner and staff, cleaned and imported years of client data, and reworked the data model when interviews showed ownership belonged on each task, not the whole application. Built role-based access, private document storage, end-to-end tests and automated deploys.
  • Presented it at Google HQ. Now leading the rollout: training staff and fixing what trips them up.
Case study →

iOS Engineer

Feb 2021 – Feb 2023
Twitter, Inc. · New York, NY
  • Engineer on the Tweet Foundation team, working on the core Tweet infrastructure every feature team in Twitter’s iOS app depended on.
  • Led completion of a major Tweet modularization milestone (legacyStatus), cutting dependencies between several teams so they could ship without waiting on each other.

Culinary Data Associate / Nutritionist

Apr 2017 – Apr 2019
Plated · New York, NY
  • Maintained the recipe and ingredient data behind production systems; managed 50 to 100 SKUs a week.

Projects

Fourth

2026
Solo build · Python, LLMs
  • Pulls CMS hospital quality data, compares each hospital’s maternal health commitments with its postpartum outcomes, scores the gap from 0 to 100 and drafts outreach. Scoring is rule-based and explainable; the LLM only writes the email. Audit logging, duplicate-send check, human approval step, 196 tests.
Case study →

AlleyCorp Relationship Intelligence

Jun 2026
Team of 3 · Next.js, TypeScript, Python, MCP
  • Built with AlleyCorp partners to show how close the firm is to people in its Deep Tech network. I owned the database schema, data enrichment and MCP server. Relationship strength comes from specific signals (co-investments, connections, events, press) that fade over time, so every rating can be explained.
Case study →

Open Source

2026
Python
  • Honcho (Plastic Labs): persistent-memory skill for LLM agents, reviewed and merged to main. Cal.com: opened a pull request improving booking-calendar ARIA labels and contrast toward WCAG 2.1 AA.
Honcho PR → Cal.com PR →

Skills

AI & LLM
LLM APIsAgent workflowsTool callingMCPRAGLLM evaluationHuman-in-the-loop reviewClaude CodeCodexCursor
Languages
PythonTypeScript/JavaScriptSQLSwiftHTML/CSS
Tools
ReactNext.jsNode.jsFastAPISupabasePostgreSQLPlaywrightGitHub ActionsVercelFirebaseGitXcode
Spoken
English (native)Russian (native)

Education

Pursuit

Sep 2025 – Present
AI-Native Software Engineering Fellowship · Long Island City, NY
  • Led rotating project teams (new team every 2 to 4 weeks): set scope, split work, reviewed code, shipped demos.

Pursuit

2019 – 2020
iOS Software Engineering Fellowship · Long Island City, NY

Brooklyn College, CUNY

2012 – 2016
B.S. in Nutritional Science · Brooklyn, NY