Tomislav – Flutter, Dart, JavaScript

The startup operates at the intersection of medicine, artificial intelligence (AI), and Health Economics and Outcomes Research (HEOR), leveraging proprietary technologies to expedite clinical research. Through the development of Artificial Clinical Intelligence (ACI), the company aims to enhance and extend human healthspan. Notably, the startup achieves significant efficiency gains by substantially reducing timelines in HEOR and Market Access research, delivering expert-validated, regulatory-compliant results at a rate that is 100 times faster than traditional approaches. This innovative approach seeks to revolutionize the pace and impact of clinical research in the healthcare industry.

• Created a complete backend solution;
• Picked tools and frameworks;
• Worked with Open AI API and Langchain;
• Worked with RAG and vector databases;
• Leading and directing a team of developers.

The product was created to automatize client intake in law firms with the help of AI. The idea of the product is to reduce costs, boost productivity and increase client onboarding capacity in law firms. Clients are also able to upload their documents and chat with them.

• Created front-end and back-end architecture;
• Coded both front-end and back-end;
• Worked with Open AI API and Langchain;
• Worked with RAG and vector databases;
• Did all the deployment and devops work.

Helping multi billion euro company to move their internal applications from outdated AngularJS framework to a modern framework.

Tomislav was part of the team with the goal of migrating outdated AngularJS applications to modern frameworks such as React and Next.js

A mobile application issued to warehouse workers with the intent to simplify the following: task distribution, navigation through the warehouse, communication between workers speaking different languages, and filling out forms. All of the mentioned was done through voice-activated commands so workers could use their hands all the time. Workers were also issued headset which communicated with the app installed on their mobile device, which was attached to their waist.

• Created custom Speech-To-Text solution from zero using Tensorflow;
• Developed mobile service app to handle communication with headset and speech recognition process;
• Designed front-end for the application in Flutter;
• Created Javascript application for human sampling speech;
• Collected voice dataset containing more than 100,000 samples;
• Made most of the architectural and design decisions;
• Used Riverpod and Freezed libraries in Flutter;
• Integrated low-level code VAD (Voice Activity Detector) in JNI and C++.

Custom mobile webshop for women's fashion with advanced filtering and attractive UI. Built for a local startup, the idea was to create a platform where small producers and sellers could sell their products with minimal hassle through the intuitive UI of the app.

• Developed front-end in Flutter as part of a team;
• Used Provider (later switched to Riverpod) for state management;
• Developed back-end in .NET Core;
• Wrote geo and full-text search queries in SQL;
• Used Firebase for cloud messaging;
• Integrated services like Google Maps;
• Used libraries like Freezed for code generation;
• Took part in the design and architectural decisions;
• Created CI/CD pipeline for deployment to Google Cloud Platform.

The project's core was to ingest as much available demographic, cultural, and social media data to create a beautiful visualization of people's purchasing habits. The heart of the solution was a machine learning model which could make correlations between data that the human eye could miss. The app can identify geographical areas of interest where your product could sell more.

• Decided on tools, front-end, and back-end technologies to be used in the project;
• Created complete front-end UI with all the capabilities using Vue.js;
• Worked on the part of the back-end in .NET Core;
• Used map tools and libraries like Mapbox, DeckGL, and Uber's H3;
• Created data visualizations using the D3 library;
• Onboarded and tutored new front-end developers;
• Used ElasticSearch to retrieve geographical data;
• Managed payment process using Stripe API.

The idea of the project was to visualize data coming from weather stations and create models for fungal and bacterial disease prediction in agriculture. Models for predictions were developed from various research papers available at the time, and the final product was distributed as a desktop, mobile, and web app to the clients.

• Created mathematical plant disease models by scraping various research papers;
• Implemented plant disease models in C#;
• Created desktop and mobile apps for visualization of weather station data and plant diseases;
• Implemented custom communication protocol between server and weather stations;
• Closely cooperated with agricultural engineers and translated their knowledge into software;
• Maintained code and added new features over the span of over three years.