I have seamlessly transitioned the engineering team of my startup to the acquiring company and integrated our error search engine into their application. With the team we achieved market leader accuracy in Rollbar’s core error grouping algorithm. To achieve this, we initiated and executed Rollbar’s AI transition: transformed legacy, hard-coded algorithms to ML. The migration improved grouping accuracy for 3,000 error types, causing 40% of crashes in the top 5 programming languages. 93% of Rollbar customers adopted the new grouping method in six months. We reimplemented Rollbar’s alerting engine in ML, reducing false alert rate by 10x, switching from managed rules to anomaly detection. I have completed a two-year AI program at Stanford University, shared the learned information with the team and applied them in our projects. I have continuously grown the company's knowledge about relevant AI technologies, including challenges, development status and achievements. I have hired, onboarded and managed a highly performing team, created an environment where they could focus on quality and think creatively.

My team automated quality control of newly synthesized compounds by processing mass spectrometer output. We developed a system to distribute chemical computations on office workstations using their idle time.

Samebug was a software startup company that built a platform to help developers to find and share solutions to software problems. It attracted over a million dollars in investment, had a team of 10 and was acquired by a San Francisco based startup. Designed and implemented the first version of Samebug from the ground up and built a talented team around it. We compiled the largest publicly available software crash database. Samebug explored how crashes are related, analyzed their root cause and collected related public knowledge from the web. We developed a novel grouping algorithm that powered the stack trace search engine. This attracted an error monitoring firm to buy our company. We had 25,000 daily visitors for over a year.

The company specified, developed, maintained and integrated or operated tailor-made software products and services for finance businesses. Provided consultancy services, helping to modernize their technology stack. Example projects: Built an online insurance broker system end-to-end, including a car insurance calculator, contract management, back-office and CRM systems. Created a high traffic online car configuration and credit application app for one of the largest Japanese car manufacturers on their main website in Hungary. Implemented a car pricing web service using Eurotax's command line software, which was used by Hungary’s leading car financing companies. Credit calculator web service supporting Toyota Financial Services Hungary website. E-Client system: online customer care directly connected to the back-office system through web services. Built an online auction site for selling used vehicles.