Skip to content

Open for full-time roles & contracts

Building distributed systems & agentic AI applications

Hi, I'm Marij — a Senior Software Engineer with 6+ years across backend, distributed systems and cloud infrastructure. Today I build multi-agent LLM applications on AWS Bedrock; before that, payments platforms and real-time telemetry at scale.

View My WorkDownload CV

About Me

From event-driven backends to agentic AI

Over the past six years I've designed and run production systems: payments microservices in Java and Scala, telemetry platforms for enterprise device fleets, and now multi-agent LLM applications on AWS Bedrock. I work across Java, Scala, Python and TypeScript, and I put as much weight on how a system is tested and evaluated as on how it's built.

What I work on:

  • Distributed systems

    Event-driven microservices, CQRS on Kafka, and zero-downtime data migrations.

  • Agentic AI applications

    Multi-agent LangGraph pipelines on AWS Bedrock, backed by DeepEval evaluation suites.

  • Security & compliance

    GDPR and CCPA crypto shredding, IAM hardening, and OAuth2 + PKCE flows.

My Journey

Work Experience

Six years across AI applications, device telemetry, payments and distributed caching.

  1. Jan 2026 – Present

    Senior Software Engineer, AI Applications

    Emumba · Extreme Networks engagement

    • Building and evaluating agentic AI applications, including a multi-agent LLM pipeline (LangGraph/LangChain on AWS Bedrock) that automates early-stage product planning — turning a raw product idea into a structured spec and then a full requirements document.
    • Designed and built a dependency-discovery agent that identifies and links internal and external dependencies for a given product initiative.
    • Built an LLM evaluation suite using DeepEval and led the testing strategy — unit, integration and smoke — across the agent pipeline.
    Show 4 more
    • Found and fixed a privilege-escalation gap in a production OpenSearch IAM policy, where a 'read-only' role's HTTP-verb-based permissions actually allowed write operations; rebuilt the Terraform-managed policy into scoped allow/deny rules.
    • Built a ~1,100-line integration test suite against a live MCP-based LLM retrieval service on AWS serverless infrastructure (Lambda, DynamoDB, OpenSearch), routing test traffic through a Kubernetes proxy to reach an IP-allowlisted endpoint.
    • Hardened platform security and infrastructure: AWS X-Ray transaction tracing, tighter IAM permissions for AgentCore, and BlackDuck SCA scanning in CI.
    • Built a batch evaluation runner for Extreme's AI Expert (EVA) chat API, and an OAuth2 Authorization Code + PKCE SSO login flow.
    • Python 3.12
    • LangGraph
    • LangChain
    • AWS Bedrock
    • Bedrock AgentCore
    • Lambda
    • DynamoDB
    • OpenSearch
    • Claude
    • Terraform
    • pytest
    • DeepEval
    • GitHub Actions
    • AWS X-Ray

  2. Oct 2024 – Dec 2025

    Senior Software Engineer, Device Monitoring Platform

    Emumba · Extreme Networks engagement

    • Led architectural decisions for a mission-critical device monitoring platform serving enterprise laptop fleets, setting technical vision and engineering standards for the distributed systems team.
    • Architected performance-critical microservices in Java 17 / Spring Boot, processing real-time telemetry from 10,000+ devices at 99.9% uptime and sub-100ms response times.
    • Designed the data architecture — PostgreSQL for relational data, Elasticsearch for search and analytics, Kafka and RabbitMQ for event streaming and async messaging — handling millions of transactions daily.
    Show 4 more
    • Built VoIP and call-quality monitoring for a C#/.NET 8 Windows telemetry service, detecting active calls via audio-session monitoring (NAudio/WASAPI) and sampling latency, jitter, packet loss and system metrics in real time, published over MQTT.
    • Implemented JWT service-to-service authentication, exposed platform services through AWS API Gateway, and stored time-series telemetry in InfluxDB.
    • Directed a company-wide migration to clean architecture and domain-driven design, establishing standards that cut technical debt by 35%.
    • Built a hybrid REST + gRPC communication architecture, reducing inter-service latency by 40%.
    • Java 17
    • Spring Boot
    • C#
    • .NET 8
    • PostgreSQL
    • Elasticsearch
    • Kafka
    • RabbitMQ
    • MQTT
    • InfluxDB
    • gRPC
    • AWS API Gateway
    • Kubernetes
    • AWS

  3. Jan 2023 – Sep 2024

    Software Engineer II

    Upstart Commerce

    • Architected high-performance eCommerce and payments microservices in Java and Scala with Akka, processing 500K+ transactions daily under strict SLAs.
    • Led the move to event-driven CQRS on Kafka, improving scalability 10x and enabling real-time processing across services.
    • Migrated a fleet of microservices from Cassandra 3 to Cassandra 4 with zero downtime and zero data loss, replaying event history — including Kafka-based inter-service events for a non-event-sourced service — to rebuild state in the new schema before cutover.
    Show 5 more
    • Implemented GDPR and CCPA compliance across event-driven microservices handling PII, using crypto shredding and Akka-GDPR to enable data erasure without breaking event-sourcing guarantees.
    • Optimized PostgreSQL performance, cutting query times by 60% through indexing, query rewriting and Redis caching.
    • Contributed fixes and features to the TypeScript/Node.js backend alongside the primary Java/Scala service work.
    • Established unit, integration and performance testing to 85% coverage, maintaining 99.99% production uptime on Kubernetes/GCP with Grafana and Prometheus.
    • Ran code reviews and mentored junior engineers on architectural patterns and Scala/Java practice.
    • Java
    • Scala
    • Akka
    • Akka-GDPR
    • Kafka
    • TypeScript
    • Node.js
    • PostgreSQL
    • Redis
    • Cassandra
    • Kubernetes
    • GCP
    • Grafana
    • Prometheus

  4. Mar 2021 – Jan 2023

    Software Development Engineer

    Upstart Commerce

    • Designed CI/CD pipelines with Bitbucket and Docker, moving the team from weekly to daily releases with automated testing gates.
    • Developed an internal performance testing platform in Node.js and React, replacing commercial tooling and enabling continuous performance monitoring.
    • Built automated testing frameworks with JUnit and TDD — 500+ tests with parallelization, cutting CI execution time by 50%.
    • Java
    • JavaScript
    • Node.js
    • ReactJS
    • JUnit
    • Cypress
    • JMeter
    • Docker

  5. Dec 2019 – Feb 2021

    QA Engineer

    Alachisoft

    • Developed automation test suites and performance testing strategies in Java, Python and C# for distributed caching systems, improving coverage by 70%.
    • Ran performance testing and system monitoring with Grafana and Prometheus for MySQL-based NCache SaaS deployments on AWS and Azure.
    • Java
    • Python
    • C#
    • MySQL
    • AWS
    • Azure
    • Grafana
    • Prometheus
    • Selenium

My Work

Featured Projects

Systems I've designed and built outside my day-to-day consulting work.

  • Aug–Sep 2026

    Natalya Rugs — shipment-to-storefront pipeline

    A Python pipeline that turns a supplier's invoice spreadsheet and a Drive folder of photos into published Shopify listings. Deterministic code handles size, price and SKU; AI handles colour, style and copy — and every AI answer has to clear a hard validator before it reaches the store.

    • Python
    • LangGraph
    • Claude
    • Pydantic
    • Typer
    • SQLite
    • Shopify GraphQL API
    • pytest
    Read the case study

  • GenAI course project

    Preference-Aware Travel RAG Assistant

    A retrieval-augmented generation assistant that gives personalized travel recommendations, using vector search to ground responses in a user's stated preferences rather than generic suggestions.

    • Python
    • RAG
    • Vector Search
    • LangChain

  • GenAI course project

    Agentic Content Curation System

    A multi-step agent pipeline built with LangChain and FastAPI that autonomously researches a topic and drafts LinkedIn post content from what it finds.

    • Python
    • LangChain
    • FastAPI
    • Agentic AI

My Toolbox

Technical Skills

The languages, platforms and practices I work with in production.

  • Languages

    • Java
    • Scala
    • Python
    • TypeScript
    • JavaScript
    • C#
    • SQL

  • Frameworks & Platforms

    • Spring Boot
    • Akka
    • .NET
    • Node.js
    • ReactJS
    • FastAPI
    • LangChain
    • LangGraph

  • AI & ML

    • Agentic AI Systems
    • RAG Pipelines
    • AWS Bedrock
    • Claude Models
    • Claude Code
    • LLM Evaluation (DeepEval)

  • Data & Messaging

    • PostgreSQL
    • MySQL
    • SQL Server
    • Kafka
    • RabbitMQ
    • Cassandra
    • Redis
    • Elasticsearch
    • DynamoDB
    • OpenSearch

  • Cloud & DevOps

    • AWS
    • GCP
    • Kubernetes
    • Docker
    • Terraform
    • CI/CD
    • GitHub Actions
    • AWS X-Ray

  • Architecture

    • Microservices
    • Event-Driven Architecture
    • CQRS
    • DDD
    • Clean Architecture
    • Performance Optimization

  • Security & Compliance

    • JWT
    • OAuth2 / PKCE
    • GDPR & CCPA
    • BlackDuck SCA
    • IAM Hardening

  • Monitoring & Tools

    • Grafana
    • Prometheus
    • OpenAPI
    • REST
    • gRPC
    • pytest

Let's build something together

Whether it's a backend that needs to scale or an AI application that needs to hold up in production, I'd like to hear about it.

Get in touchView LinkedIn