Dominik Kaukinen

I architected and deployed a comprehensive robotic process automation (RPA) solution to overhaul fragmented, manual workflows for insurance quoting and policy lifecycle management. By engineering a robust, automated pipeline that seamlessly integrated with core API platforms, the system autonomously executed data extraction, retrieved pricing, and generated initial policy documents. It also successfully streamlined complex lifecycle events such as renewals, endorsements, and cancellations. It incorporated strict validation logic and fail-safes to ensure data integrity. Ultimately, this deployment drastically reduced processing times, eliminated manual data entry errors, and accelerated quote-to-bind ratios, enabling the processing of a significantly higher volume of policies without scaling administrative overhead.

I architected a low-latency incremental character-level speech recognition system using bidirectional RNNs optimized with CTC loss. I made this to learn the state of the art (at the time) for machine learning architectures for modeling sequence-to-sequence tasks.