Backend Engineer — distributed systems & production AI
I build production AI systems for corporate finance on top of Spring Boot microservices on AWS — and I care about making LLM agents safe to run on top of real money.
whoami
I'm a backend engineer focused on distributed systems and cloud-native architectures. I work across Java, Python, Go, Docker, Kubernetes and AWS, building services and AI-driven applications that run in production.
At Mendel, I'm a Semi-Senior Backend Engineer on the AI platform. I grew our MCP server from 15 to 69 tools — listed in Anthropic's Claude connector directory and used by 250+ users at 60+ customer companies — and I own the WhatsApp expense assistant, a Spring Boot service plus a Python LangGraph agent., I build the AI platform behind intelligent expense management — a Spring Boot microservices ecosystem on AWS EKS. I own the WhatsApp conversational experience end to end: a Java orchestration layer and a Python AI agent for natural-language understanding and tool-based workflows.
Much of my work is making LLM-initiated actions safe on financial data: a multi-tenant OAuth gateway with MFA, SSO and delegated tokens for AI agents, idempotent writes with Redis locks, deterministic guardrails and an evaluation rig that replays conversations. Previously at Accenture I built multi-agent AI systems with RAG pipelines, and I teach Microservices Architecture at ITBA. to production, exposing expense-audit and travel-approval tools to AI assistants like Claude — with an OAuth 2.0 authorization layer (Dynamic Client Registration, refresh tokens), delegated SSO and SSE streaming transport. Previously at Accenture I architected multi-agent AI systems with RAG pipelines, and I teach Microservices Architecture at ITBA.
focus[]
Intelligent automation and agent workflows that hold up under real production traffic.
Model Context Protocol servers that expose real tools to AI assistants, safely and with auth.
Services tuned for latency, throughput and cost — where the numbers actually matter.
Containerized workloads on Kubernetes, with the CI/CD and observability to keep them healthy.
Microservices that stay resilient as they scale, communicating over well-defined contracts.
git log --experience
Promoted from Junior in Jul 2026. Grew our MCP server from 15 to 69 tools (listed in Anthropic's Claude connector directory; 250+ users at 60+ companies in Q3), shipped 22 tools in under two weeks for enterprise customers, and designed a multi-tenant OAuth gateway with MFA, Microsoft SSO and delegated tokens for AI agents. I own the WhatsApp expense assistant (Spring Boot + Python LangGraph): completed expense reports grew +166% quarter over quarter in Q3, and 88% of the receipts it processes end in a completed report.. Owner of the WhatsApp conversational experience (Java BFF + Python AI agent) and of our production MCP server with OAuth 2.0, delegated SSO and SSE streaming.
Teaching a class of 40+ students to design and build microservice-based systems — lectures and hands-on workshops with Docker, Kubernetes and CI/CD, focused on deploying, scaling and observing distributed systems., cloud-native tooling and modern communication patterns. Focused on deploying, scaling and observing distributed systems.
Built a multi-agent AI system on a customized n8n fork that routes user queries to specialized agents — designing agent workflows, integrating RAG pipelines and optimizing LLM response quality. Earlier, as an intern, worked across RAG/GraphRAG, prompt engineering and fine-tuning to improve enterprise-chatbot accuracy, cost and performance.
education --list
Graduated with an Honors Diploma for academic excellence..
work[]
Monitoring and operational automation for Docker environments — my ITBA final project.
Full-stack review app with a RESTful HATEOAS API — Spring + Hibernate backend, Angular front.
stack --list
contact