Mkrtych O.

I help enterprise teams ship production AI applications — RAG systems, LLM-powered analytics, agentic workflows, and intelligent data pipelines that turn unstructured information into business outcomes. What I build: Retrieval-Augmented Generation (RAG) systems over enterprise documents, knowledge bases, and operational data LLM applications using OpenAI, Google Vertex AI, xAI Grok, and self-hosted Ollama for privacy-sensitive workloads Vector search infrastructure with pgvector, plus integration with Pinecone, Chroma, and Weaviate LangChain and LlamaIndex pipelines, prompt engineering, evaluation, and guardrails AI-driven analytics over IoT and time-series data (InfluxDB, Cassandra, Kafka) Java/Spring Boot and Python backends, AWS deployment (Lambda, IoT Core, Bedrock-ready) Why clients hire me: Senior Software Engineer at Deloitte Consulting, currently delivering on Walmart, BMO, Elevance, and American Home Insurance engagements — I understand what production at enterprise scale actually requires $100K+ earned on Upwork across 21 jobs, with consistent 5-star ratings on AI and complex full-stack work BS in Electrical Engineering and Computer Science, Cal Poly Pomona US-based (Glendale, CA), available for collaboration in US business hours What makes me different: Most AI engineers stop at the API. I have spent over a decade shipping firmware, embedded systems, and IoT platforms — which means I can bridge LLMs, RAG, and agentic AI all the way down to the device, the sensor, and the time-series store. If your AI initiative touches operational data, industrial systems, or connected devices, that crossover is hard to find anywhere else. Typical engagements: Build a RAG-based knowledge assistant over your internal docs, contracts, or SOPs Stand up an LLM-powered analytics layer on top of existing data warehouses or IoT streams Architect and harden an AI feature that already works in prototype but is not production-safe Audit an existing LangChain or vector DB stack and re-engineer it for cost, latency, and accuracy If you are scoping something in any of these areas, send me an invite or message me directly. I respond fastest to specific, well-scoped requests; I am happy to do a free 15-minute scoping call before either of us commits.