Hire the Best Neo4j Developers

Hi I'm Dario, 27 years old programmer and architecture designer from Zagreb, Croatia. Why am I different from others? Well first of all, I am always here, ready to help you bring your ideas to life. You don't need just another coder. You need a technical partner who understands architecture, business logic, and how to ship scalable products fast. With almost a decade of experience building complex systems for high-level clients like the European Commission and various Ministries, I bring "big tech" quality to your project. I specialize in modernizing workflows—combining rock-solid Software Engineering principles with cutting-edge AI Tools to deliver results 10x faster than traditional developers. HOW I CAN HELP YOU: 🤖 AI & Automation Solutions • Building Custom AI Chatbots (RAG) that talk to your specific data. • Python Automation Scripts to replace manual data entry/Excel work. • Integrating OpenAI/Claude APIs into existing apps. 🏗️ Senior Full-Stack Development • Frontend: React.js, Next.js, Vue.js, TypeScript, Tailwind CSS. • Backend: Python (FastAPI, Flask), Node.js, SQL (Postgres/Supabase). • Architecture: Designing scalable MVPs that don't break as you grow. 🔍 QA & Code Audits • Cross-platform testing (Web & Mobile) to ensure bug-free launches. • Codebase Audits: I review your current code to fix performance issues and security risks. WHY WORK WITH ME? • Education: MSc in Software Engineering (FER). • Speed: I use modern "Vibe Coding" workflows (Cursor, AI Agents) for rapid delivery. • Communication: I explain complex tech in plain English. Whether you need a full build, an automation script to save you time, or a consultation to validate your strategy—I am ready to help. Let’s build something future-proof.

I'm a Senior IT Consultant and Engineer with 6+ years of experience building scalable web, mobile, cloud, and AI-powered applications. I specialize in developing production-grade software that combines modern full-stack engineering with the latest advancements in Generative AI and Large Language Models (LLMs). My expertise spans AI-native application development using OpenAI, Claude, Gemini, DeepSeek, and other frontier models to build intelligent products, AI agents, conversational experiences, content generation systems, document processing pipelines, and workflow automation. I enjoy "vibe coding"—rapidly transforming ideas into production-ready software by combining AI-assisted development with strong software engineering principles. I have hands-on experience designing and integrating: - AI Agents & Agentic Workflows - LLM-powered Applications - Retrieval-Augmented Generation (RAG) - Model Context Protocol (MCP) - Prompt Engineering & AI Automation - OpenAI, Claude, Gemini, DeepSeek, and Video Generation APIs - REST & GraphQL APIs - Cloud-native backend services and scalable microservices Beyond AI, I have extensive experience building high-performance applications using Node.js, TypeScript, React, Next.js, React Native, NestJS, Express.js, Spring Boot, Apache Kafka, GraphQL, PostgreSQL, MongoDB, Neo4j, and AWS. Throughout my career, I have led engineering teams, architected scalable backend systems, optimized cloud infrastructure, and delivered enterprise-grade software for high-volume businesses. I believe the best software comes from combining clean architecture, strong engineering practices, and practical AI solutions that solve real business problems. As a Neo4j Certified Professional, I also bring expertise in graph databases, enabling intelligent relationship modeling, knowledge graphs, recommendation systems, and advanced data analytics. I'm passionate about building the next generation of AI-powered products and always excited to collaborate on innovative software that pushes the boundaries of what's possible.

Rebuilding your SaaS backend after it fails under real user load can be costly and take a long time. I assist SaaS founders and CTOs in getting the architecture right from the start, so you won’t have to rebuild in 18 months. I have over 9 years of experience with .NET, Azure, and modern architecture patterns. I’ve completed more than 4,700 hours on Upwork, achieving a 100% Job Success Score and earning Top Rated Plus status. Here’s where I deliver results: - Multi-tenant SaaS architecture, with role-based security, Stripe billing, CI/CD, and observability integrated from day one, not added later. - AI integration for existing .NET apps using Azure OpenAI, Semantic Kernel, and RAG (Azure AI Search or Pinecone) without altering your current stack. - Cross-platform mobile with .NET MAUI — shared C# codebase for iOS and Android, native push notifications, offline sync, and App Store/Play deployment, backed by the same .NET API layer. - Legacy modernization, upgrading from .NET Framework to .NET 8/10 using Blazor and Azure App Services. - Scalable APIs and microservices using CQRS, MediatR, YARP/Ocelot gateways, RabbitMQ/Azure Service Bus, Docker, and Kubernetes, applying event-driven design where it adds value. - Third-party and enterprise integrations, including OAuth2, Xero, Autodesk ACC, Microsoft Graph, and HubSpot, with complete control of authentication, token handling, and data mapping. My tech stack includes ASP.NET Core, C#, Azure, Blazor, PostgreSQL, SQL Server, MongoDB, Docker, Semantic Kernel, MediatR, CQRS, RabbitMQ, EF Core, and Dapper. I take on a limited number of projects so I can remain involved with architecture decisions, not just implementation. If that sounds like what you need, let’s talk. Message me with details about your project, and I’ll let you know honestly if I’m the right fit.

I build AI agents, RAG pipelines, and voice AI that survive production — not just the demo. Free 30-min consultation: I map where an LLM actually saves you money. Useful even if we never work together. Recent results: NLP system covering 1M+ questions across 1,500+ classes. Document OCR at 7 seconds per document. On-device voice cloning from a 15-second sample. WHAT I DO • AI Agents & LLM Integration — chatbots, autonomous agents, tool use, workflow automation with LangChain, OpenAI, n8n. Cost-control by design: model routing, caching, usage limits. • RAG & Semantic Search — retrieval pipelines over your PDFs, contracts, tickets, and knowledge bases, with citations and evaluation, not just embeddings. • Voice AI — speech recognition, text-to-speech, voice cloning, personal voice activity detection (PVAD), offline/on-device voice assistants. • Document AI — OCR and data extraction from court documents, receipts, and multi-page PDF packages; classification, routing, and API/Kubernetes deployment. • Computer Vision — object detection and tracking (YOLO), segmentation, depth estimation. SELECTED RESULTS • Banking chatbot NLP: detection and clustering of unknown questions — tested on ~1M real questions, 1,500+ classes (Python, PyTorch, BERT, HDBSCAN). • Court document processing: OCR + attribute extraction + routing at 7 sec/document on a 4-core CPU; 100-page PDF packages fully automated (Tesseract, OpenCV, Kubernetes, REST API). • Voice cloning on Android: local TTS from a 10–15 second voice sample, running fully on-device — no cloud GPU (PocketTTS, voice embeddings). • Offline warehouse voice assistant: hands-free order picking via headset, works without internet (VOSK speech recognition + TTS). • Personal Voice Activity Detection: 100+ model experiments, 1+ TB of audio data — isolating a target speaker in noisy multi-speaker calls (PyTorch, DEMUCS). • Receipt automation (CheckMate): OCR + categorization + auto-generated accounting reports — several times faster than manual processing. • Passenger counting from bus video: 92% accuracy (YOLOv8, Jetson). HOW I WORK • Starting a NEW AI project? I'll do a free planning session / architecture sketch first, so you know scope and cost before spending budget. • Have an EXISTING product? I'll do a free architecture or code review and show you the bottlenecks — yours to keep either way. • Milestones you approve before payment releases. Weekly updates. Documented, tested code. BACKGROUND 15+ years in software engineering. 8,000+ hours delivered on Upwork as part of the Singularis Programming team (Top Rated Plus, 100% Job Success), including an Upwork Enterprise client. I state my role on team projects explicitly — what you see in my portfolio is what I actually did. Tech: Python, PyTorch, TensorFlow, LangChain, OpenAI API, RAG, LLM fine-tuning, n8n, Docker, Kubernetes, AWS, PostgreSQL, SQL, MLOps, Tesseract OCR, OpenCV, YOLO, VOSK, C#, .NET. Available today — message me and we can have a 15-minute call to see if I can help. 1. Artificial Intelligence 2. LLM 3. Retrieval Augmented Generation 4. LangChain 5. Natural Language Processing 6. Semantic Search 7. OpenAI 8. Machine Learning 9. Deep Learning 10. PyTorch 11. n8n 12. Python 13. MLOps 14. Computer Vision 15. Data Extraction 16. Docker 17. Amazon Web Services 18. Data Analysis 19. SQL 20. Data Visualization

Agents don't fail because the model is bad. They fail because the tools they're handed are unnavigable, the retrieval is unmeasured, and nothing catches a regression before the client does. I build the layer underneath: production MCP servers (Model Context Protocol), RAG pipelines with measured accuracy, and the eval harnesses that keep both honest. 🏆 Top Rated Plus · $350K+ earned · 8,000+ hours billed ━━━━━━━━━━━━━━━━━━━━━━ RECENT PRODUCTION RESULTS 🔌 Built and shipped a production MCP server (FastMCP, Python) that gives a client's analysts direct agent access to their own domain data — questions that used to need an engineer now get answered in the chat window. It runs in production behind an agent service on AWS Fargate over stdio, and in Claude Desktop and Claude Code. I designed the tool surface for progressive discovery — broad list, then filter and count, then drill down — so models navigate 15K+ records without blowing their context, and built credential-gated tool registration with a read-only-by-design data layer so pointing an agent at live data is safe. 850+ tests, and an eval harness I run across model versions before shipping changes, so a model upgrade can't silently break agent behaviour. 🤝 Wrote the agent that drives it, too — a project-level Claude Code subagent with a curated tool allowlist, in-prompt gates derived from real production failures, and a cost-tier routing matrix that picks the cheapest transport likely to work. Building the tool surface and the agent that consumes it is a different skill from wiring up one API. 🤖 Built a RAG extraction service (FastAPI + Celery + Pinecone, two-tier model routing with a per-model cost estimator) turning messy documents into structured data across ~11K projects from 11 registry sources. Measured on a golden dataset, accuracy went from 33% to 91% F1 on one extraction task and 43% to 84% on another — the difference between a pipeline nobody trusted and one the team runs unattended. Every output traces back to its source document. 🧠 Built a second production RAG system on a medical knowledge graph (FastAPI, Neo4j, MongoDB) with character-level span citations, so a disputed claim takes seconds to check rather than an afternoon. Application-layer tenant isolation with dedicated tests proving no cross-tenant leakage. 2,700+ tests; I wrote roughly two-thirds of the codebase. 🌍 Built and operate a scraping platform covering 20 sources behind enterprise anti-bot protection — Cloudflare-class WAFs and Incapsula, a rotating datacenter proxy pool plus a residential tier for the hardest targets, and fallback transport chains that step up only when they have to. 225K+ documents collected to date. 50+ scheduled pipelines, around 30 of them daily, with per-source error recovery and alerting. ━━━━━━━━━━━━━━━━━━━━━━ WHAT I DO ✔️ AI integration & agent infrastructure — MCP (Model Context Protocol) servers with FastMCP, tool surfaces designed for how models actually search, Claude Desktop and Claude Code integrations, custom Claude Code subagents, prompt engineering, eval harnesses ✔️ LLM & RAG backends — Anthropic/OpenAI/Gemini APIs, Pinecone, vector search with RRF fusion, structured extraction from messy documents, measured accuracy against golden datasets, cost routing that sends the easy 80% to cheap models ✔️ Web scraping & data extraction — Playwright, Selenium, ZenRows; resilient access to protected sources, proxy management, scheduled fleets via Celery, PDF/Excel/Word extraction, normalization into clean schemas ✔️ API development — FastAPI, Flask, Django; auth, rate limiting, background jobs, clean documentation ✔️ Distributed systems — Celery, RabbitMQ, Redis; retries, idempotency, fault tolerance under real load. Redis caching at 85-95% hit rate, typically 10-50x faster responses ✔️ Production ops — Docker, AWS, PostgreSQL/MongoDB/Neo4j, 110+ zero-downtime migrations on a single project, CI-gated test suites running 2,500-4,900 tests on my largest systems ━━━━━━━━━━━━━━━━━━━━━━ HOW I WORK ✅ I own systems end-to-end: architecture → implementation → deployment → monitoring → handover docs. ✅ I'll tell you when an LLM is the wrong tool — and what to use instead. Cheaper for both of us than finding out in week three. ✅ Failures surface where you'll see them: Prometheus/AlertManager into Slack with per-alert templates, Grafana and Loki for dashboards and logs, scheduled digests and a daily data-feed tripwire. I get paged, not you. ✅ Most of my $350K+ comes from repeat clients and multi-year engagements. ━━━━━━━━━━━━━━━━━━━━━━ 📍 UK-based (GMT/BST) If you need agent tooling that real models can navigate, an LLM pipeline whose accuracy you can actually check, or scraping infrastructure that survives contact with real anti-bot systems — send me a couple of lines about your project and I'll tell you straight away whether I'm the right fit.

Most “AI agent developers” build a chatbot wrapped around an API call. I build stateful, production-grade multi-agent systems that plan tasks, call tools, write and execute code in sandboxed environments, recover from failures, and remain fully traceable through LangSmith. Not prototypes. Not demos. I build systems that support real concurrent users, run unattended in production, and provide the traces needed to understand exactly what happened. That is the gap I fill: The space between “I got GPT to answer questions” and “this AI system runs reliably in production.” WHAT I’VE SHIPPED — PRODUCTION, NOT DEMOS - Multi-agent HR automation using LangGraph, GPT-4o, and Azure. Reduced recruiter processing time by 70% while supporting 1,500 concurrent users. - Enterprise RAG assistant using Gemini, hybrid retrieval, GCP, and Kubernetes. Reduced query resolution time by 60% and supported up to 1,800 concurrent users. - AI coding agent with sandboxed execution. Built a LangGraph agent that plans, writes, runs, tests, and self-corrects code inside isolated Docker/E2B-style environments. - Legal RAG pipeline for a 500+ document corpus. Custom chunking and LangChain evaluations improved retrieval accuracy by 40% compared with baseline vector search. - Agentic travel platform built with LangGraph Cloud, GCP, and LangSmith. Automated the complete workflow with no manual intervention. - Visual AI agent builder similar to Gumloop, with configurable workflows, tool integrations, and production LangSmith tracing. - AI email automation platform built with FastAPI, NLP classification, and RAG. Eliminated 80% of manual support-ticket triage. Read this section first. Everything below provides technical context. The systems above are the proof. WHAT I BUILD AI Agents & LLM Orchestration - LangGraph multi-agent systems - Stateful planning and execution loops - Conditional branching and parallel sub-agents - Tool calling with retries and error recovery - GPT-4o, Claude, Gemini, and open-source model routing - Human-in-the-loop approval workflows - Long-running and asynchronous agent tasks - LoRA and QLoRA fine-tuning Coding Agents & Sandboxed Execution I build coding agents that do more than generate code. They write, execute, test, inspect errors, and revise their solution using real execution feedback. - Code generation and execution - stdout, logs, stack-trace, and test-result analysis - Automatic error recovery and self-correction - Docker and E2B-style sandbox environments - Timeouts, resource limits, network isolation, and controlled tool access Enterprise RAG - Hybrid vector and keyword retrieval - Neo4j knowledge graphs for multi-hop queries - Pinecone, FAISS, Weaviate, and PostgreSQL vector search - Custom document chunking and parsing - Cross-encoder reranking - LangChain evaluation pipelines - Measurable accuracy and relevance improvements Python Backend Development - Async Python and FastAPI - REST APIs, WebSockets, and SSE streaming - PostgreSQL, MongoDB, Redis, and Supabase - HubSpot, Salesforce, Stripe, and custom API integrations - Dockerized microservices - Kubernetes-ready architecture - Background workers and GitHub Actions CI/CD Observability, MLOps & Cloud Deployment - LangSmith tracing, evaluations, and debugging - Token, latency, and cost monitoring - GCP, Cloud Run, GKE, Vertex AI, and LangGraph Cloud - Azure AI, AKS, and Azure Functions - Model deployment and versioning - Production testing and reliability monitoring CORE STACK / KEYWORDS AI Agent Development, LLM Engineering, Generative AI, Agentic AI, Multi-Agent Systems, LangGraph, LangChain, LlamaIndex, LangSmith, OpenAI GPT-4o, Claude, Gemini, RAG, Enterprise RAG, Pinecone, FAISS, Weaviate, Neo4j, Docker, E2B, Python, FastAPI, PostgreSQL, MongoDB, Redis, GCP, Azure, Cloud Run, Kubernetes, GitHub Actions, MLOps, React, and TypeScript. WHAT YOU GET Every delivery includes production-ready architecture, documentation, automated tests, secure tool and sandbox execution, deployment support, LangSmith observability, failure recovery, and post-launch support. I don’t just hand over code. I hand over AI systems that run, recover, scale, and can be debugged when something goes wrong. NOT THE RIGHT FIT FOR ManyChat flows, Zapier replacements, Make automations, simple chatbot integrations, basic API connections, no-code AI tools, or quick proof-of-concept scripts. My engagements are Python-first, architecture-focused, and built for production. LET’S TALK Message me with your project scope. I’ll tell you clearly whether I can solve it, what architecture I recommend, and how I would approach the build.