Language: English | Русский
Python / AI Product Engineer | Local-first systems | FastAPI | Automation
Portfolio • Repositories • Telegram • LeetCode
I build privacy-conscious AI products, deterministic automation systems, and reliable Python backends. My repositories focus on runnable setup, explicit service boundaries, local data ownership, and evidence-backed engineering claims.
| Repository | Why it matters | Stack |
|---|---|---|
| vasya_ai | Local-first desktop assistant with voice/text workflows, tasks, calendar, memory, integrations, and a protected local API | Python, FastAPI, SQLite, Ollama |
| college_auto_schedule | College scheduling system with guarded Excel imports, versioned local data, readiness checks, and CP-SAT generation | Python, FastAPI, SQLite, OR-Tools |
| taskflow_orchestrator | Backend-heavy async task orchestration with durable state, retries, dead-letter handling, and operational metrics | FastAPI, PostgreSQL, Redis, Celery |
| career-radar | Evidence-first local career intelligence with deterministic matching and a private Opportunity Inbox | Python, SQLite, FastAPI |
| aboutme | Bilingual macOS-style portfolio with GitHub project sync, Markdown blog pipeline, previews, and Pages delivery | React, GitHub Pages |
Local-first AI vasya_ai, vasya_ai_landing
Backend systems taskflow_orchestrator, career-radar
Automation / EdTech college_auto_schedule, attendance_bot
Frontend portfolio aboutme, advenubi-landing, js-marvel
Telegram bots tajnyj_ded_bot, python-tasks-taskbot, finance_manager_bot
- Local-first AI assistant: voice UX, memory, task/calendar flows, integration boundaries, and API protection.
- Scheduling automation: validated Excel ingestion, versioned SQLite storage, readiness reports, solver benchmarks, and Windows alpha packaging.
- Backend architecture: idempotent APIs, background workers, retries, dead-letter states, metrics, Docker Compose, and CI checks.
- Portfolio system: auto-synced GitHub projects, bilingual content, Obsidian-based posts, project previews, and deployed static delivery.
- Local-first AI assistants for voice, tasks, notes, memory, and daily automation.
- FastAPI services with typed contracts, persistent state, background jobs, and operational checks.
- Deterministic EdTech and workflow automation built around validated source data.
- Local browser interfaces that make complex technical workflows inspectable and usable.
- Keep private data local and make every external integration explicit.
- Separate product claims from plans, prototypes, and private evidence.
- Prefer narrow end-to-end increments with focused regression tests.
- Document setup, architecture, limitations, and verification so another engineer can inspect the result.
- aboutme v0.1.0: macOS-style portfolio, GitHub project sync, Markdown blog, RU/EN content.
- vasya_ai v0.6.0: tagged milestone for the local-first assistant.
- tajnyj_ded_bot v0.1.0: documented Telegram bot release.
- Hardening local-first assistants and private-data boundaries.
- Advancing deterministic scheduling, career intelligence, and workflow automation.
- Improving reproducible setup, CI gates, release evidence, and developer documentation across the portfolio.



