Boran Oktay D.

AI Automation Engineer | Claude, n8n & RAG Agents in Production

$260,000 earned on Upwork across 125 completed projects, 5,229 hours billed, 100% Job Success Score, Top Rated Plus. I build AI automation that a business actually runs on: agents, RAG pipelines and end to end workflows wired into the tools a team already uses. LangChain and LangGraph orchestration in Python and FastAPI, Claude and OpenAI integrations, vector search on Pinecone and ChromaDB, with n8n, Make or Zapier as the no-code layer when that is the right call. Most AI work dies in the gap between a model that works in a demo and a process that works on a Monday morning. Closing that gap is the whole job. Clients I have built for include Trukkin (logistics), LGFG Fashion House (fashion), Ultrassure (insurance), Genvision (climate tech), Bowery Boost (performance marketing), Moor Marketing (lead generation) and Ta3swim (ecommerce). Istanbul, UTC+3. I overlap 09:00 to 13:00 US Eastern every weekday and answer inside 4 hours outside that window. You work with me directly. I write the code myself. 𝗪𝗛𝗔𝗧 𝗜 𝗕𝗨𝗜𝗟𝗗 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬 𝐚𝐧𝐝 𝐋𝐋𝐌 𝐒𝐲𝐬𝐭𝐞𝐦𝐬 𝐢𝐧 𝐏𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 Single and multi-agent systems, LangChain and LangGraph orchestration, Claude and OpenAI integrations, MCP servers and connectors, function calling, human approval gates wherever a mistake is expensive, plus evaluation and guardrails so you can tell whether the thing is actually working. 𝐅𝐢𝐱𝐢𝐧𝐠 𝐚𝐧𝐝 𝐇𝐚𝐫𝐝𝐞𝐧𝐢𝐧𝐠 𝐖𝐡𝐚𝐭 𝐘𝐨𝐮 𝐀𝐥𝐫𝐞𝐚𝐝𝐲 𝐇𝐚𝐯𝐞 A workflow that ran fine for months and then stopped. An agent that answers confidently and is wrong. A prototype that works on one laptop with the API keys sitting in a file. I diagnose it, fix it, and leave it on a footing where it keeps running: logging, error handling, retries, version control, permissions. This is most of my current work. Three of my active contracts are improving and extending systems that were already in production, including a weekly ML agents program running since February 2025. 𝐑𝐀𝐆 𝐎𝐯𝐞𝐫 𝐘𝐨𝐮𝐫 𝐎𝐰𝐧 𝐃𝐨𝐜𝐮𝐦𝐞𝐧𝐭𝐬 Retrieval that stays accurate as the document set grows, with source citations so any answer can be checked. Chunking, embeddings, Pinecone, ChromaDB, GraphRAG and evaluation. When a first attempt disappoints, the problem is almost always retrieval, not the model. Also support and sales assistants grounded in your own data, with escalation to a human when confidence is low. 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐒𝐲𝐬𝐭𝐞𝐦 𝐈𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐢𝐨𝐧 Two way syncs between systems that were never designed to talk to each other. APIs, webhooks, and the payload changes that quietly break things overnight. CRM integration including ZOHO, internal tools, email, databases and data flows across your stack. 𝐒𝐚𝐚𝐒 𝐚𝐧𝐝 𝐌𝐕𝐏 𝐃𝐞𝐯𝐞𝐥𝐨𝐩𝐦𝐞𝐧𝐭 End to end AI powered SaaS builds. Python and FastAPI backends, authentication, billing, dashboards, cloud deployment on AWS, GCP or Azure. From a blank repository to a live product with paying users. 𝐃𝐚𝐭𝐚 𝐏𝐢𝐩𝐞𝐥𝐢𝐧𝐞𝐬 𝐚𝐧𝐝 𝐒𝐜𝐫𝐚𝐩𝐢𝐧𝐠 Large scale web scraping, ETL, cleaning and enrichment. The unglamorous layer that decides whether an AI system is actually useful. 𝗥𝗘𝗦𝗨𝗟𝗧𝗦 𝗕𝗬 𝗜𝗡𝗗𝗨𝗦𝗧𝗥𝗬 𝐑𝐞𝐚𝐥 𝐄𝐬𝐭𝐚𝐭𝐞: 30+ automation projects (Zillow, CoStar), AI powered house price calculators, ZOHO CRM integrations. 𝐃𝐢𝐠𝐢𝐭𝐚𝐥 𝐌𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠: AI SEO researchers, content generators, YouTube analytics agents, full blog production pipelines. 𝐇𝐞𝐚𝐥𝐭𝐡𝐜𝐚𝐫𝐞 𝐚𝐧𝐝 𝐒𝐚𝐚𝐒: AI chatbots, personalized email systems, lead generation platforms taken from zero to live products. 𝐒𝐚𝐥𝐞𝐬 𝐚𝐧𝐝 𝐆𝐨 𝐭𝐨 𝐌𝐚𝐫𝐤𝐞𝐭: AI sales agents, lead research and qualification, outreach personalization at scale. 𝗖𝗨𝗥𝗥𝗘𝗡𝗧𝗟𝗬 𝗕𝗨𝗜𝗟𝗗𝗜𝗡𝗚 A weekly ML agents program on a single production system, running since February 2025. A RAG orchestration and prompt engineering layer on top of an existing LLM product. An AI content pipeline in daily production use. Long engagements rather than one off scripts, which is where the 100% Job Success Score comes from. 𝗧𝗘𝗖𝗛 𝗦𝗧𝗔𝗖𝗞 𝐋𝐋𝐌𝐬 𝐚𝐧𝐝 𝐀𝐠𝐞𝐧𝐭𝐬: Claude, OpenAI GPT, Gemini, MCP servers and connectors, agentic and multi-agent architectures, LangChain, LangGraph, function calling, evaluation and guardrails 𝐑𝐀𝐆 𝐚𝐧𝐝 𝐃𝐚𝐭𝐚: RAG pipelines, GraphRAG, embeddings, vector databases, PostgreSQL, ETL, large scale scraping 𝐁𝐚𝐜𝐤𝐞𝐧𝐝: Python, FastAPI, Node.js, REST APIs, webhooks, Docker 𝐀𝐮𝐭𝐨𝐦𝐚𝐭𝐢𝐨𝐧: n8n, Make, Zapier, ZOHO CRM, third party API integration 𝐂𝐥𝐨𝐮𝐝: AWS, GCP, Azure 𝗛𝗢𝗪 𝗜 𝗪𝗢𝗥𝗞 I ask about the business before I write code. I tell you when something should not be automated. I document what I build in plain English, so your team can own it after I am gone. Send me the process you want covered and I will tell you what it can own, what stays with your team, and how long it takes.