Denis K.

I design and develop AI and machine learning systems, including AI agents, multi-agent applications, RAG pipelines, MCP servers, and business automation. I use Python, LangGraph, LangChain, n8n, and LLM APIs to connect AI applications with business data, external tools, and production services. My work spans system architecture, development, deployment, and optimization. I focus on reliable execution, secure integrations, and reducing avoidable latency, token usage, and infrastructure costs. AI Agents & Multi-Agent Systems I build tool-using agents and coordinated multi-agent systems, including orchestration, delegation, state persistence, memory, context management, and human-in-the-loop approvals. I work with LangGraph, LangChain, Model Context Protocol (MCP), agent-to-agent (A2A) patterns, and Dapr-based orchestration, choosing the architecture that fits the task rather than adding agents unnecessarily. Production AI Engineering & Durable Execution I design long-running AI workflows that recover from failures without repeating completed work. My focus includes durable execution, checkpoints, event-driven processing, retries, idempotency, external events, human approval, observability, and access control. My work draws on durable workflow patterns used by Dapr Workflows and Diagrid Catalyst, including preserving progress across interruptions, validating resumed state, and preventing duplicate side effects. LLM Engineering & RAG I develop AI applications that retrieve information from documents, websites, databases, and internal knowledge sources. This includes ingestion, chunking, embeddings, semantic and hybrid search, reranking, retrieval evaluation, grounding, and context optimization. I also design LLM integrations around model selection, structured responses, evaluation, caching, latency, and cost control. I work with vector databases and search technologies such as Milvus, Zilliz, Qdrant, Weaviate, Chroma, and FAISS, as well as other retrieval architectures. I select tools based on data volume, relevance, latency, and deployment needs rather than committing to a single database. MCP Servers & API Integration I develop MCP servers, tool interfaces, REST APIs, and backend services that let AI systems retrieve data and take authorized actions. My work includes authentication, input validation, structured outputs, pagination, rate limits, retries, error handling, and secure integrations. I use Python, FastAPI, JavaScript, TypeScript, and Node.js. n8n Automation & Business Systems I build agentic and rules-based workflows connecting LLMs, APIs, databases, and business applications. These include approvals, customer support operations, document processing, notifications, data synchronization, and scheduled tasks. I integrate tools such as n8n, NocoDB, Shopify, Twilio, and SendGrid, and build custom integrations when off-the-shelf nodes are insufficient. Web Scraping & Data Engineering I develop crawlers, browser automations, extraction pipelines, and ETL processes for analytics, AI applications, and RAG ingestion. I work with Playwright, Puppeteer, Selenium, Python scraping frameworks, browser session management, proxy integration, dynamic sites, and structured data processing. I also assess anti-bot compatibility, extraction reliability, speed, and infrastructure cost. Machine Learning & Software Engineering My background covers predictive modeling, classification, NLP, sentiment analysis, recommendations, computer vision, and model evaluation. I develop backends and full-stack applications using Python, Django, FastAPI, React, JavaScript/TypeScript, SQL databases, and Docker. I also work with deployment, tracing, monitoring, Kubernetes, Grafana, and AWS services. Technical Background & Publications I hold a BSc in Computer Science. I've published technical articles for Diagrid, ZenRows, Groundcover, Zilliz, Twilio, The New Stack, Strapi, and MakeUseOf. These articles cover technologies I also build with, including AI agent architecture, durable workflows, RAG, MCP, web automation, machine learning, distributed systems, and observability. If you need an AI agent, RAG application, MCP server, business automation, or help improving an existing AI system, send me your requirements. I can support architecture, implementation, integration, troubleshooting, and optimization.