Yehor Z.

Senior Python Backend Engineer and AI Developer with 7+ years of production experience, 8,500+ hours on Upwork, $200K+ earned, and a 100% Job Success Score (Top Rated Plus). I specialize in AI development, AI agent development, data automation, and backend engineering. My work spans LLM-powered agents, RAG pipelines, distributed REST APIs, data extraction systems, and cloud-native infrastructure on AWS. Every project I take on ends with a system that runs in production — properly architected, monitored, and ready to grow with the business.

Technical stack that I can provide for your project:
🏗️ Backend Engineering & API Development: FastAPI and Django are my primary tools — async RESTful APIs with RBAC, billing hooks, WebSockets, background workers, and OpenAPI docs, built the right way from day one. I integrate with third-party services via OAuth, webhooks, and event-driven consumers, and I design software architecture with layered separation and dependency injection — code the next engineer will actually want to work with. Database-wise I work with Flask, Kivy, PostgreSQL, MongoDB, Redis, DynamoDB, and SQLite, with proper SQL optimization, migration management, and query tuning. I've built microservice systems over gRPC, REST, Kafka, SQS, RabbitMQ, and Celery, and real-time pipelines with WebSockets, Redis pub/sub, and live data feeds.

🤖 AI & LLM Development: I build LLM agents with multi-turn state using LangChain and LangGraph — the kind that run in production on Telegram, Discord, and WhatsApp. For knowledge retrieval I work with RAG pipelines backed by Qdrant and LLaMA-Index, with real semantic search over internal documents. Prompt engineering is a proper workflow here — evaluation frameworks, LangSmith observability, structured output validation. I integrate agentic workflows through API (OpenAI, Claude, Anthropic, Gemini) with full output control.

⚙️ Data Automation & Extraction: Data automation pipelines with Dagster, dbt, Pandas, PyArrow, and Polars — reliable, scheduled, and debuggable. I extract structured content from PDFs, Office documents, HTML, and APIs using BeautifulSoup, PyMuPDF, python-docx, and openpyxl. Web scraping at scale with async Python, proxy rotation, rate limiting, and retry logic that actually holds up under real conditions. For unstructured sources I apply NLP and ML classification models to produce clean, usable output from data.

☁️ Cloud & Infrastructure: I work with the full AWS stack in production — S3, SQS, Lambda, Athena, DynamoDB, EKS, Kinesis, Bedrock, CloudFront, also Docker, Kubernetes (Helm + ArgoCD), and Terraform for reproducible infrastructure as code. Observability built in Datadog APM, OpenTelemetry, Sentry, and vector database management with Qdrant and AWS OpenSearch for semantic search and RAG retrieval.