OPEN TO WORK / PROJECTS
Python Backend Developer · AI/LLM

Ilya Minin

I am a Python Backend Developer with over a year of experience and an AI/LLM focus. I design APIs and data models with FastAPI, work with PostgreSQL, Redis and task queues, write tests, and support production services.

I independently built and launched Kantano, a collaborative task management platform with a LangGraph / LangChain AI assistant: LLM APIs, tool calling, persistent state, and user approval for data changes.

Open to remote work, on-site or hybrid roles in Moscow and Saint Petersburg, and relocation.

Tech Stack

// Backend

  • Python
  • FastAPI
  • asyncio
  • SQLAlchemy
  • Pydantic
  • REST APIs
  • WebSocket
  • SSE
  • pytest

// AI/LLM

  • LangGraph
  • LangChain
  • AI agents
  • LLM APIs
  • Tool calling
  • Human-in-the-loop

// Data & Messaging

  • PostgreSQL
  • Redis
  • RabbitMQ
  • Celery
  • Alembic

// Infrastructure

  • Docker
  • Docker Compose
  • GitHub Actions
  • Caddy
  • Linux
  • Git
  • S3

// Observability

  • Grafana Stack
  • OpenTelemetry
  • Sentry

// Additional

  • Vue 3
  • TypeScript
  • OpenAI Codex
  • Claude Code

Experience

May 2025 - June 2026

Backend Developer

Overleng-IT LLC

Automated academic paper compliance service operating according to GOST requirements.

  • Built REST APIs and business logic with FastAPI for key thesis and term-paper workflows, enabling automated compliance checks against GOST requirements.
  • Designed PostgreSQL models and implemented SQLAlchemy queries for new service functionality.
  • Moved academic-paper checks to Celery/RabbitMQ: the API now returns a task ID in under 1 second instead of waiting 30-60 seconds for processing.
  • Automated typed TypeScript client generation from OpenAPI, reducing backend-frontend API contract synchronization from 20-25 minutes of manual edits to a single command.
  • Covered delivered functionality with pytest unit and integration tests against an isolated database, maintaining approximately 85% overall backend coverage.
  • Established diagnostics using Grafana metrics and logs, reducing the time to localize common errors and bottlenecks from several hours to 15-20 minutes.
  • Improved backend stability by fixing defects, refactoring code, and reviewing changes; maintained the Docker environment and CI/CD.
Python FastAPI PostgreSQL Redis RabbitMQ Celery SQLAlchemy pytest Docker CI/CD Grafana

Projects

kantano.ru · github.com/Ilm1n/Kantano

Kantano - AI-powered Task Manager

Production collaboration platform with an AI assistant

Independently designed, developed, and launched a collaborative Kanban service; I maintain it in production.

  • Email verification, JWT/OAuth, roles and invitations; board updates and presence over WebSocket for real-time collaboration.
  • LangGraph/LangChain AI assistant with 14 typed tools for tasks, columns and tags: tool calling and human-in-the-loop controls, with data changes requiring user approval and permission checks.
  • Dependent multi-step plans, persistent state in PostgreSQL, run recovery, and SSE streaming to preserve context across requests.
  • Google Gemini, GigaChat and OpenAI-compatible APIs with fallback across cloud models.
  • Background email delivery with RabbitMQ/Celery, files in S3 and Redis caching: repeated board loads take 4 SQL queries instead of 10.
  • 300+ backend tests with pytest and 86% coverage; linting, type checking, and TypeScript client generation from OpenAPI.
  • CI/CD, Grafana Stack, Sentry and S3 database backups; monitoring of LLM latency, tool calls, fallback events and token usage.
FastAPI LangGraph LangChain PostgreSQL Redis RabbitMQ Celery WebSocket SSE Docker Vue 3 TypeScript S3 Grafana Stack
Code on GitHub ↗

Education

2026

Academy of Urban Environment Management, Urban Planning and Printing

Diploma in Information Systems and Programming (09.02.07)

Secondary vocational education

Languages

Russian - Native English - B2 (technical communication and documentation)