Open to opportunities

Ivan Medvedev

@ivanmedvedev

Message

Full-Stack AI engineer building production AI integrations, RAG, agents, and real-time voice systems.

Vietnam

Message

What I'm looking for

I’m looking to build and ship production AI agent and RAG systems—especially real-time, cost-efficient workflows—with strong engineering standards around reliability, observability, and clean architecture.

I’m a full-stack engineer with 5+ years of commercial work, now focused on production AI integration. I’ve shipped four AI systems running live: real-time voice interviews on OpenAI Realtime API with Gemini scoring; a blockchain RAG assistant; a Telegram community agent using a Qwen-to-Claude cascade to cut LLM cost ~80%; and a 24/7 client-acquisition agent driving native Chrome sessions through Playwright MCP and Claude Code.

At WorkorAI, I shipped 70% of the codebase (143 of 204 commits) and built end-to-end real-time voice interviews, Gemini embedding + evaluation pipelines (LLM-as-judge with rubrics), and a centralized AI client with structured outputs and prompt-injection guards. I also lead production AI work across validator-information products, running a Claude Code multi-agent build/review pipeline on a nightly schedule—while keeping critical errors down to 0.3% and improving indexing speed by 25% across Cosmos, Namada, Ethereum, Solana, and Polkadot.

Experience

Work history, roles, and key accomplishments

Current

Core Engineer

Current

WorkorAI

Nov 2025 - Present (7 months)

Built an AI talent platform from scratch and shipped 70% of the codebase (143/204 commits) to production. Implemented real-time voice interviews with OpenAI Realtime API and Gemini, plus LLM-as-judge evaluation and candidate–role matching.

OpenAI API Google Gemini RAG LLM As Judge Zod Schema Validation PostgreSQL

Current

Full-Stack & AI Engineer

Current

Citizen Web3

Dec 2024 - Present (1 year 6 months)

Led production AI work across validator-information products, including a Telegram community agent where Qwen-to-Claude routing cut Claude API cost by ~80%. Built RAG assistants over indexed podcast transcripts and live on-chain state, plus a production multi-agent build/review pipeline and a 24/7 Chrome-based acquisition agent.

Qwen Anthropic: Claude Vercel AI SDK RAG Playwright MCP Docker MCP Servers Multi Agent Pipelines

Full-Stack Engineer

FinTechAssociation

Jan 2024 - Dec 2024 (11 months)

Designed FastAPI + Vue3 architecture and shipped a Cosmos redelegation service from scratch in five months. Built a Cosmos network analytics web service (ETL and key-metric visualization) and prepared it for deployment in three months.

fastAPI Vue3 ETL Redis Docker Nginx JavaScript

Full-Stack Engineer & Data Analyst

Innovation Agro Group Ltd.

Mar 2021 - Dec 2023 (2 years 9 months)

Designed React + Django REST architecture and shipped a corporate platform from scratch in six months. Built 10+ internal services and an R&D analytics tool that sped research analysis 3x, supporting four patents with structured validated data.

React Django REST Framework Python Pandas NumPy Chart.Js Data & Analytics JavaScript