Software & application engineering

Complex systems. Practical thinking. Working software.

I'm Dino Lozina, founder of Akritech. I build applications end to end — from the interface and backend to integrations, deployment, and production operations. My work spans conversational AI, distributed platforms, and developer tools.

Based in Split, Croatia · Working remotely

Game-inspired pixel art of a coastal engineering studio at dusk, with glowing computer screens overlooking the sea and distant islands.
The workshop, after hours.
ArchitectureApplication engineeringConversational AIProduction ownership

01 / Selected work

Built to do
the real work.

Workflow orchestration, voice infrastructure, secure integrations, and cloud platforms — selected work from design through production.

Workflow diagram: a call passes to a waiting timer and then review, with a recovery path back to the waiting stage.

Workflow orchestration

Workflows that recover from failure

Founded a service for multi-step workflows connecting calls, messages, waits, and human review. Built conditional routing, retry and resume, bulk recovery, and operational monitoring.

The outcome

Work can resume after failures, with automatic retries when a provider recovers and visibility into tasks that need attention.

  • Python
  • FastAPI
  • Cloud Tasks
  • Pub/Sub
Voice diagram: an audio waveform and a server platform exchange signals in both directions.

Voice & messaging

The infrastructure behind the conversation

Extracted telephony from a larger application into a regional service owning phone-number registries, number pools, and outbound calling. Built number-management tools and integrated voice and SMS into the wider platform.

The outcome

A dedicated production service with clear ownership of calling infrastructure and the tools to manage it.

  • Python
  • LiveKit
  • React
  • TypeScript
Integration diagram: an event passes through an approval gate before delivery.

Secure integrations

Connecting AI to real operational work

Delivered an after-hours callback integration from design through production, including secure event delivery, human approval, recording and transcript exports, and knowledge-base synchronization.

The outcome

Conversation outputs move into downstream systems through an auditable workflow with a human review gate.

  • Outbox pattern
  • Mutual TLS
  • FastAPI
  • React
Platform diagram: build, run, and observe stages form a pipeline, with monitoring feedback returning to the build stage.

Distributed systems & developer experience

Cloud platforms developers can build on

Designed and implemented serverless execution and JavaScript offloading services. Built SDKs, command-line tools, API libraries, and documentation, and led distributed engineering teams through architecture and delivery.

The outcome

Experience across both the distributed platform and the interfaces developers use to work with it.

  • Rust
  • WebAssembly
  • TypeScript
  • Kubernetes

02 / How I can help

A thinking partner.
A hands-on engineer.

End-to-end application engineering: interfaces, backend services, integrations, deployment, and production support.

01

End-to-end application engineering

Build the complete application, from the user interface and APIs to data, integrations, and deployment. Carry it through release and into production support.

02

Backend & platform engineering

Design and build services, APIs, event-driven workflows, and integrations. Turn a new product requirement or a growing monolith into a system with clear boundaries.

03

Conversational AI systems

Connect voice agents and LLMs to telephony, messaging, workflow orchestration, and human review. Handle the scheduling, state, and recovery around the conversation.

04

Architecture & technical leadership

Work through architecture decisions, write actionable technical proposals, review code and APIs, and help engineering teams deliver changes across services.

05

Production reliability

Build recovery paths, observability, and controlled rollouts into the system. Improve consistency across services and make failures easier to understand and resolve.

The details are part of the work. Consistent state across services. Clear access boundaries. Recovery paths and observability. These are part of the implementation, from the start.

03 / The person behind the code

Hi, I’m Dino.

From machines that move
to systems that talk.

Industrial automationDeveloper toolingCloud platformsConversational AI

I'm a software engineer based in Split, Croatia. I started in industrial automation, programming machines and robotic systems, and moved into developer tools, distributed cloud platforms, and conversational AI.

That path shapes how I work: understand the system, make its behavior explicit, and take responsibility for what happens after it ships. I combine hands-on implementation with architecture, technical leadership, and code review.

Alongside engineering, I've taught robotics and advanced control systems and authored educational books on robotics. Akritech is my software engineering consultancy, established in 2022.

Explore my engineering journey, robotics books, and interests on my personal website .

Tools I work with

  • Python & FastAPI
  • Rust & WebAssembly
  • TypeScript & React
  • GCP & Kubernetes
  • Event-driven systems
  • Voice & LLM integration
Find me on GitHub

04 / Let’s build something useful

What are you
working on?

Tell me about the system you're building, the problem you're facing, and where you need help.

dino@akritech.eu