Innokentiy Vetrov

Innokentiy Vetrov

Efficiency Engineer & AI Systems Architect

"The Engineer of the "Soulful Machine""

Paris, France inno@vetrov.fr

Executive Summary

AI Research Engineer bridging audiovisual production and industrial automation. Focused on efficient inference, multimodal retrieval, and reliable AI infrastructure powering premium creative workflows.

Core Competencies

Advanced Computer Vision

CLIP, DiT, video generation pipelines

High-Performance DevOps

Multi-cloud (AWS, GCP, OVH) for AI inference

Neuromorphic R&D

SNN + generative video research

Full-Stack Automation

OpenAI/Anthropic API integration

Professional Experience

Efficiency Engineer & Co-Founder

The Metamodern Company 2022–Present
  • AI visual pipeline architecture and inference reliability
  • Custom workflows for multimodal generation (Python, ComfyUI)
  • Applied R&D on vision and generative models

Process & Data Manager

Ooshot 2022–Present
  • Multi‑cloud infrastructure (AWS, GCP, OVH, Scaleway) for AI inference
  • DevOps pipelines (GitLab CI, DockerHub) for rapid deployment
  • Model operations and fine‑tuning for creative production
  • Automation pipelines (Make, Zapier) with OpenAI/Anthropic APIs
  • On‑prem compute setup for private AI experimentation

Full-Stack Developer (Oosearch Project)

Ooshot 2022
  • Hybrid text–image search engine (React, Node.js, PostgreSQL, CLIP, Pinecone)
  • AWS migration for inference‑heavy workloads

Data Scientist Apprentice

Ooshot 2020–2022
  • Data quality maintenance, backups, and CRM migrations
  • Automation tools (Monday.com, custom scripts) for ops efficiency

Lead Video Editor & IT Ops

Clapnclip 2018–2020
  • Creative video production (Premiere Pro, After Effects)
  • Post‑production automation tooling
  • Internal technical infrastructure maintenance
  • Supervised junior editors and coordinated production

Education

2022

Master's AI & Management

IA School, Paris

2017

Scientific Baccalaureate (CS)

Lycée Isaac Newton

Current Research

CIFRE PhD candidate: Generative Video Inference & Neuromorphic Computing