Lazar P.

I’m an AI Agent Engineer, Automation Specialist, and Full-Stack Developer with a deep passion for embedding intelligence into real-world web and mobile products and eliminating manual business workflows through automation. My core focus is designing, building, and deploying AI-driven systems—from LLM-powered agents and voice assistants to data-aware automation pipelines that integrate seamlessly with production applications. I work extensively with large language models (LLMs), retrieval-augmented generation (RAG), vector databases, embeddings, tool-calling frameworks, and agent orchestration patterns to create systems that reason, act, and scale reliably. On the backend, I build high-performance APIs and services using FastAPI, Flask, Node.js, and Django, with strong experience in async architectures, background workers, and event-driven systems. I regularly work with PostgreSQL, MySQL, Redis, MongoDB, and Neo4j, and design robust data layers using SQLAlchemy, Prisma, and ORM/DAO patterns. For AI workloads, I integrate OpenAI, Anthropic, Ollama, ElevenLabs, and custom inference pipelines, alongside Qdrant, Pinecone, FAISS, and hybrid search setups. On the frontend, I build modern, scalable interfaces using React, Next.js, Vite, Tailwind CSS, and TypeScript, with a strong emphasis on UX for AI-powered tools (real-time streaming, agent state visualization, conversation flows, and admin dashboards). For mobile, I work with React Native and cross-platform architectures to bring AI features directly into user-facing apps. Automation is a major part of my work. I design end-to-end business automation systems using [login to view URL], n8n, webhooks, cron jobs, message queues, and direct API integrations. This includes automating CRM workflows, email systems, call routing, lead qualification, reporting, and data synchronization across platforms like Twilio, Slack, Google Sheets, Stripe, Shopify, and internal tools. I’m highly experienced in AI voice agents and real-time systems, integrating Twilio, speech-to-text (STT), text-to-speech (TTS), streaming APIs, and conversational state management to build production-ready voice and chat agents. I also work on computer-vision and multimodal pipelines, combining vision models, tracking, and sensor data for advanced use cases. From an infrastructure and MLOps perspective, I deploy and scale systems using Docker, Linux servers, cloud VPS, CI/CD pipelines, and GPU-accelerated environments. I care deeply about observability, logging, error handling, security, and cost optimization in AI systems. Above all, I love turning business problems into automated, intelligent systems—reducing operational friction, increasing efficiency, and building products where AI is not a feature, but a core capability. If you need AI agents, automation pipelines, or full-stack systems that actually work in production, that’s exactly what I build.