Lead Backend & AI Engineer — Arian System
Architecting an on-premises RAG and MCP-enabled AI assistant across commercial ERP products.
Senior Software & AI Architect
Engineering Experience. Building Intelligent Futures.
I build enterprise-grade software and intelligent AI systems that turn complex business challenges into scalable solutions.
Eighteen yearsbuilding the software companies keep their books in. Lately, teaching them to think.
Years of Engineering Experience
Years at Motogen
Enterprise AI Engineering
Real enterprise systems, not toy demos.
On-premises RAG, Local LLM, LangGraph and MCP integration.
Modular architecture, enterprise APIs and performance optimization.
Enterprise manufacturing systems, ERP, MIS, DSS and BI.
Event-driven architecture, Redis and real-time sensor workflows.
RAG · LLM · LangGraph · MCP · Multi-Agent Systems
Designing on-premises and hybrid AI pipelines with measurable accuracy, strict data privacy, and live ERP integration.
DDD · Clean Architecture · Distributed Systems · Microservices
Structuring complex business domains into maintainable, decoupled systems built to evolve over decades.
C# · .NET · Python · FastAPI · API Design
High-throughput APIs, event-driven workflows, and database performance tuning for mission-critical workloads.
Technical Strategy · Team Leadership · Engineering Management
Aligning engineering execution with business outcomes, mentoring senior developers, and leading digital transformations.
18+ years of shipping enterprise systems.
Architecting an on-premises RAG and MCP-enabled AI assistant across commercial ERP products.
Guiding production readiness, vector ingestion pipelines, and local LLM evaluation strategies.
Built a low-latency sensor-data workflow engine and event-driven microservices for digital twins.
Architected a modular ERP platform on ASP.NET Core and SQL Server, reducing API latency by 75%.
Led legacy modernization from a monolith into 9 integrated ERP, MIS, DSS, and BI modules for industrial manufacturing, cutting executive reporting from 12 hours to under 1 hour.
Architecture should tame complexity, not become the source of it.
Technology is only valuable when it solves a real, measurable business problem.
An intelligent system must be evaluable, observable, secure, and maintainable.
Great enterprise systems are designed for change, clarity, and decades of evolution.
AI has made writing code cheap, but verifying it is as expensive as ever. Six architectural guardrails to settle before adopting AI at speed.
Why naive vector search fails on ERP documentation, and how hybrid retrieval, evaluation suites, and MCP solve it.
Lessons from decomposing monolithic industrial software without halting daily operations.
Let's build something solid.