Kaspars Polis

• LLM Reinforcement Learning & Post-Training: Applied LoRA and GRPO reinforcement learning to post-train a multimodal foundation model for production personalization workflows. Built training datasets, evaluation benchmarks, and post-training workflows for continuous model improvement. • Trained a design decomposition model using object detection, Pix2Pix, SPADE, and CycleGAN on 65K+ designs • Build a ResNet color profile classifier with 97% accuracy and about 3,000 daily predictions

* AI/ML Urban City Traffic Light controlling with SUMO using Reinforcement learning - PoC Developing Agent-based Intelligent Control System Algorithm research Model training Technologies: SUMO, TraCI, TF, AWS DeepLens, AWS VPC , EMQX, Python, three.js

System maintenance and feature implementation Issue investigation, CI/CD Pipeline automatization Delivery, support and fixing daily issues, support lead role Technologies: Jira, Azure, Jenkins, Atlassian, Chef, Linux, Selenium, Splunk. Methodologies: DevOps, Scrum, Kanban, Agile

Design of systems aimed at recognizing keywords in the Latvian language Audio signal classification, research into algorithms, maintenance of infrastructure Computer network monitoring, administrations and introduction of automation Firewall, gateway administration, IT projects supervision Following project deadlines Implement new features and fix issues across the system, and find best algorithms. Technologies:Python Tornado API, WEKA, Matlab, Sphinx4, Xen, Docker, PostgresQL, EL, Sphinx