Ilya Roger

Ilya Roger

Software Developer & Systems Specialist @Vention

Building AI-powered search infrastructure for large-scale podcast indexing: @PodZeus

Specialized in:
Architect and deploy distributed systems · Run and manage inference on NVIDIA A30s · Build containerized worker pipelines in Go · Handle high-throughput media processing with ffmpeg & RabbitMQ · Optimize Linux environments for performance, reliability, and cost.


Tech I enjoy:
Go · TypeScript · Bash · PostgreSQL · Docker · Nginx · ffmpeg · Ubuntu · Pinecone

Work experience

Vention - Minsk, Belarus

Fullstack Developer | January 2024 - present

Building internal AI proof-of-concepts, automation workflows, and semantic data tools for client-facing products. Also provisioned NVIDIA A30 inference infrastructure and led practical AI workshops across the company.

KaiserKoenig - Hamburg, Germany

Frontend Developer | April 2022 - Oct 2023

Built React applications, testing workflows, and internal tools for media and subscription products. Work included Storybook/Chromatic CI, Jest/Cypress coverage, a Vue-to-React migration, and AI-assisted research tooling for Der Spiegel.

iTechArt - Minsk, Belarus

Frontend Developer | Jan 2022 - Feb 2022

Joined after relocating from Frankfurt to Minsk, but the role ended almost immediately due to the war in Ukraine and resulting sanctions. Relocated back to Germany when the company could no longer operate locally.

BioEticGoods - Ticino, Switzerland

Frontend Developer | Feb 2020 - Aug 2021

Built and shipped frontend features with React, TypeScript, and Next.js while coordinating directly with design and product stakeholders. Focused on performance, code quality, Lighthouse improvements, and weekly client delivery.

Sagos - Basel, Switzerland

Frontend Developer | Jan 2019 - Feb 2020

Created frontend systems and reusable UI foundations with React, Storybook, ESLint, and Prettier. Improved performance through lazy loading, asset optimization, and Lighthouse-focused delivery.

Freelance Project - Basel, Switzerland

Frontend Developer | Feb 2019 - Oct 2019

Delivered a frontend project across design, implementation, SEO, and performance work. Used Figma, Zeplin, Tailwind CSS, Jest, and GitHub Actions to keep the product shippable and maintainable.

Projects

DispatchPing - Field Maintenance Scheduling Optimizer

June 2026 - Present

DispatchPing turns field-service planning from spreadsheet work into optimized weekly schedules. It matches jobs to technicians by location, certification, time window, and SLA, then produces route maps, timelines, and field-ready exports.

PodZeus - Podcast Search & Intelligence Platform

December 2025 - Present

PodZeus makes podcast archives searchable by exact moments, not just episode titles. It uses transcription, timestamps, lexical search, and semantic search to help creators and advertisers find relevant clips fast.

ClipZeus - YouTube Analysis for Clippers

April 2025 - November 2025

ClipZeus analyzed YouTube videos for short-form clippers, helping identify high-potential moments from long-form content. The project was paused after YouTube API and bulk-download restrictions made video ingestion at scale impractical.

ChefGPT - Recipe Generator

Jan 2023 - Feb 2023

ChefGPT generated country-specific recipe collections with OpenAI, turning prompt output into a browsable recipe site. It was a small experiment in AI-generated content, formatting, and export-ready recipe pages.

Education

SAE Institute - Frankfurt am Main, Germany

B.Sc. Games Programming | Mar 2019 - Jun 2021

  • Delivered multiple game projects from conceptualization to production within tight deadlines, employing agile methodologies and versioning tools like Plastic SCM for efficient workflow
  • Engineered a custom graphics engine using C++, incorporating advanced rendering techniques like shader programming and texture mapping
  • Mastered industry-standard principles in game design and development, gaining hands-on experience with frameworks like Unity and Unreal Engine

Skills

AI

  • WhisperX (ASR pipeline)
  • Pinecone (vector search)
  • GPU provisioning (A30 / A100 setup, CUDA tuning)
  • Model serving (inference endpoints, Nebius / local GPU nodes)
  • Token optimization & batching for inference cost control
  • Embedding generation + retrieval architecture

Backend

  • Go (high-concurrency workers, API services)
  • PostgreSQL
  • RabbitMQ (job queues)
  • ffmpeg / yt-dlp (media processing)
  • Rate limiting, retries, and queue-based scaling

Infra

  • AWS
  • Linux (Ubuntu optimization for GPU + I/O)
  • Docker, Docker Compose
  • Nginx (reverse proxy, secure download delivery)
  • WebShare (residential proxy rotation)
  • Bash scripting, systemd, Supervisor
  • Hetzner servers (dedicated compute nodes)
  • Basic observability (logs, uptime checks, queue metrics)

Languages

English

C2

German

C2 - Native

Russian

C1 - Native

Belarusian

A2