Designed and built software for a GPU-based cloud capture service (~200 Windows servers) producing ~1M clips/day for 2.5M+ MAU and 15+ fully edited clips/s at peak, with fully autonomous operation and centralized logging/monitoring.

Built Python/C++ tooling to capture gameplay from demos in both landscape and native portrait, manipulating games to record only key moments and inject personalized content.

Delivered first-to-market CS2 VODs at scale and cut cost-per-clip by up to 3× via multi-title concurrency, CPU affinity pinning, process prioritization and OS-level tuning.

Built and own an FFmpeg + Python video-editing pipeline adding intros/outros, watermarks, slow motion, overlays and user-specific branding for studio-quality clips.

Implemented high-throughput, templated thumbnail generation on AWS Lambda in Python for all clips.

Developed proprietary C++ reverse-engineering–based tech to replace in-engine assets on demand for brand placement and advanced clip customization, powering 10+ high-engagement sponsored campaigns.

Improved visual quality and UX via engine-level fixes, including CS2 portrait HUD adjustments and depth-buffer/FX fixes for CS2 smoke rendering in HLAE/ReShade.

Built a Dota 2 demo parser in Java and Node.js/TypeScript backend pipelines to feed personalization and metadata into capture and editing.

Led multiple R&D and prototyping efforts to solve complex technical problems and evaluate new product ideas, acting as a key voice in technical feasibility and direction.

Designed and built a fully autonomous Raspberry Pi–based camera system for harsh outdoor conditions, handling unstable power and network environments

Developed Python services for automatic camera control, image capture, local storage and upload to a central server, managed via Linux system services and automated deployment

Implemented support for multiple camera types (GigE industrial cameras via GStreamer, vendor SDK–based cameras, Raspberry Pi cameras), with remote configuration and control for each

Built image processing and analysis pipelines using OpenCV (cv2, histograms) and dynamic camera parameter tuning (exposure, gain, white balance) to reliably capture night-sky imagery

Developed safety and protection logic, including robotic sun shields and temperature-based protections that prevent operation outside safe ranges and control heating systems automatically

Added monitoring and alerting (including email alerts on failures) to keep remote installations observable and maintainable with minimal manual intervention

Led technical discovery for a university equipment management and sharing system, running requirements workshops, mapping user stories and coordinating stakeholders.

Prototyped the full user journey in Figma based on business needs and future-user feedback, producing a development-ready action plan, waterfall breakdown and cost estimates.

Drove pre-development analysis for a government document-management/CMS overhaul, evaluating Nuxeo against performance, customization, workflow and secure-sharing requirements.

Built a Java demo plugin for Nuxeo to showcase low-level extensibility and microservices-style integration options.

Contributed to an OCR-based invoice processing system (Java Spring Boot, SQL, email automation), investigating bug reports, performing root-cause analysis and stabilizing production behavior.

Became the primary liaison between the government client and the dev team, translating technical findings into clear language and aligning expectations on system behavior and limitations.

Automated and built KPI reporting tools and dashboards, giving the branch manager and leadership timely, clear visibility into performance and operations.

Consolidated data sources and standardized reports to reduce manual work and ad-hoc analysis.

Owned regulatory reporting for donation spending and transparency, handling sensitive data on sponsored healthcare professionals and non-monetary benefits.

Ensured accurate, on-time public transparency reporting in line with AbbVie corporate standards and local regulations.

Managed and tracked donation projects, ensuring proper documentation, compliance and on-time delivery.

Led and supported CRM/CLM platforms: configuration, user support, troubleshooting and data quality management.

Acted as on-site technical support for the affiliate, resolving hardware/software issues and enabling smooth day-to-day operations.

Supported rollout of new software in the affiliate, including testing, user onboarding and ongoing support.

Designed and delivered MS Excel trainings for all Baltic branches, equipping employees with more efficient tools and workflows to work smarter.

Built an Excel VBA solution that automatically generated weekly driver reports (deliveries, revenue, historical performance)

Consolidated multiple data sources and automated revenue/stat calculations, replacing manual spreadsheet work

Saved roughly one full day of work per week by eliminating manual report creation and data checks