I learn by building

Below are some of my projects. As a practitioner, when I learn about a new technology, I don't apply it to client work right away. First, I build at least one real product with it to see if it's good enough for a production environment. All of the below are my personal solo builds. Every new skill must be polished to increase quality, security and stability of the outcome.

Otelic - observability built on OpenTelemetry

The Idea

To give developers access to production logs and traces without needing direct server SSH access. It's an observability platform for any application using OpenTelemetry libraries, designed to centralize metrics, traces, and logs in one place.

The Execution

A platform built on top of the OpenTelemetry open-source standard. It receives telemetry data, allowing developers to browse and analyze application behavior through a clean web interface.

Tech Stack

TypeScript React React Query Tailwind CSS MongoDB ClickHouse

Screenshots

Trafikito - monitor output of any Linux command

The Idea

At Trafikito you can monitor the output of any command and execute an API call when stuff happens.

The Execution

React, Node.js, and POSIX shell scripts. It runs on AWS (EC2 and DynamoDB) with global infrastructure on three edges for fast loads worldwide. I hired a freelancer to double-check my shell code.

Tech Stack

DynamoDB Node.js React MongoDB AWS

Screenshots

Oisie - event driven marketing automation

The Idea

To build a marketing automation tool like Klaviyo. It allows non-technical team members to create powerful, event-driven email sequences and segment users based on their in-app behavior.

The Execution

A system where developers define events, and marketers build automation flows. It can send targeted emails or trigger webhooks based on what users do (or don't do).

Tech Stack

TypeScript Node.js React Next.js MongoDB

Screenshots

BNR - AI driven support and community tooling

The Idea

The AI content machine. At its foundation are full legal documents used to answer questions for my local market (Lithuania) about bendrijų sodų bendrijos and DSNB bendrijos. I built an admin panel that uses leading AI models to create full content, cross-link by SEO standards, and cover topic clusters. People can ask their own questions and get AI-powered answers. New articles, updates, and many other levers drive SEO traffic. An ongoing experiment in content marketing and AI-powered community building.

The Execution

Admin-driven content pipeline: AI generates and updates articles, maintains internal linking and topic clusters, and powers a Q&A experience so every visitor can get a tailored answer. The system is built to scale content and SEO in a niche legal domain.

Tech Stack

TypeScript React React Query Tailwind CSS MongoDB Node.js Leading AI models k3s cluster

Screenshots

Leadpurify

The Idea

To create a simple, fast tool for validating email lists. Email lists degrade over time, and sending to invalid addresses hurts deliverability. This service ensures a list is clean before a campaign is sent.

The Execution

A service where users upload a CSV of contacts. The backend parses data, extracts social links, performs live SMTP checks to verify each email address, and returns a clean, verified file.

Tech Stack

TypeScript React Node.js MongoDB API Integrations

A Note on My Process

I built all of these products entirely by myself-from backend architecture and API design to the UI/UX and final implementation. They are my proving ground for exploring new technologies, marketing approaches, and development patterns.