Valts Darznieks

It was a joint effort between multiple organizations, and I was lucky enough to have the chance to work on the project as the lead front-end engineer.

The project was completed in three weeks and had more than one million active users in the first few days.

I helped design and build an AI-driven learning platform that automatically generates interactive educational content for children. The core of the system lets you define structured scenes, rules, and success criteria, which are then prompt-engineered into dynamic, AI-generated instructional scripts, so new lessons can be produced without hand-authoring every step. I worked across the full stack and contributed to end-to-end pipelines covering content generation, interaction logic, and post-session analysis, which fed back into improving future sessions. The stack was Next.js, React, TypeScript, and Tailwind on the front end, with PostgreSQL and Prisma for data, and LangChain plus Langfuse for large language model (LLM) orchestration and observability. My focus was reliable, structured LLM output and making the generated content consistent and safe enough to put in front of kids.

I designed and built the 1st social game in Latvia, which reached 200,000 active users in its first quarter. As a front-end and full-stack developer, I owned the gameplay client and the integration with the social network platform it ran on, building for fast growth and a large concurrent audience from day one. I also developed a popular friend-quiz game on the Hungarian social network Iwiw.hu and prototyped several 3D Flash social games while exploring new mechanics and formats. This was early, high-traffic social gaming, long before it was a mature field. It taught me how to ship engaging, performant front-end experiences that scale to hundreds of thousands of users on the web.