Skip to content
About me

Vladimir Kharlashko

Architect of Autonomous Multi-Agent Systems

Serbia (Belgrade) · @VladimirKharlashko

I architect systems where AI agents operate as an engineering team: with contracts, quorum, evaluation, and kill gates. 15 years in IT, the last 2 focused on AI engineering.

I started as a DevOps engineer, grew into a team lead, then architect, then CTO. At every level I saw the same thing: people drowning in routine decisions while critical ones were made by a single person with no cross-check. At PGK I governed 150+ IT initiatives and delivered ~$5.5M in economic effect. At PARMA I cut overdue projects from 55% to 15% in six months and drove AI adoption across 500+ staff. The pattern kept repeating: the core problem is never code — it's decision-making. That's why today I build systems where decisions are made by panels of multiple AI models with quorum, not by one person and not by one model. Not because AI is better than people — but because disagree-and-commit between models catches errors that a single agent misses.

Path

Three eras, one thread

2011–2019DevOps Engineer → Team LeadFullstack, microservices, CI/CD
2019–2024Architect & CTO~$5.5M effect, 150+ initiatives
2024–nowAI Architect22 projects, 92 integrations, consensus engine, 553 tests
Testimonials

What leaders say

One of those rare leaders who doesn't just 'manage projects,' but builds systems where change becomes part of the culture and teams drive continuous progress.

Dmitrii Riabov, Enterprise ArchitectPARMA Technologies Group

Under his leadership, our team not only achieved its goals but consistently exceeded expectations by 120% across all targets.

Simon Vazeykin, Head of IT Operational EfficiencyPGK / First Freight Company

A unique blend of technical expertise, strategic vision, and a proactive approach, making him instrumental in driving impactful transformations.

Roman Golik, Product LeadPGK Digital
Numbers

By the numbers

~$5.5M
economic effect — PGK, Russia's largest private freight operator
55→15%
overdue in 6 months — PARMA, IT holding
500+
staff adopted AI tools, −50% routine work
20
cases with deep dive and interactive demos
150+
IT initiatives governed in Arch Committee
15
years: DevOps → TL → Architect → CTO → AI Architect
Leadership

Not just code — teams and strategy

10+ years in leadership: from Chief Technical Expert to CTO. Hiring, mentorship, architecture committees.

2024–2026

CTO / Chief Architect

PARMA Technologies · Insurance

  • P&L: 1.1B RUB payroll, 400+ systems, revenue target 4B RUB
  • AI strategy across 5 departments: presale, delivery, legal, HR, sales
  • ROI-based buy-in via Strategic Development Committee (CEO + founder)
2020–2024

Director of Architecture

PGK / First Freight Company · Logistics, railway ($1.5B revenue)

  • Built architecture function from scratch, 100+ people in org perimeter
  • Introduced design review process and architecture committee
  • Monolith to microservices migration: 15+ services
2018–2020

Chief Architect / VP R&D

Raiffeisenbank · Fintech, banking

  • Bank-wide architecture committee governance
  • Design review for all major projects (50+ reviews)
  • Hiring and mentoring: 10+ architects and leads
2017–2018

Chief Technical Expert

Moscow Exchange (MOEX) · Capital markets

  • Architecture of trading and clearing systems
  • High-load systems (millions of transactions/day)
AI Experience

Career path

2024–now

AI/LLM Architect & Builder

Full cycle: from multi-agent system architecture to production deployment. 22 projects, 92 integrations, 4 servers.

2019–2024

Architect & Engineering Manager

Architecture and engineering team management. ERP, fintech, e-commerce. Up to 22 people in a flat structure.

2011–2019

Software Engineer → Team Lead

Fullstack development, Python/JS. Growth from engineer to team lead. Microservices, CI/CD, DevOps.

Domains

10+ AI application domains

AI agents are not a single tool, but an approach that scales to any domain.

Multi-Agent Systems

92 services, 10 pipeline stages, DAG code generation

Legal Analysis

AI panel for settlement strategy, NDA review

Cold B2B Sales

OSINT → DeepSeek → personalization → SMTP

Content @ Scale

10 languages × 13 rubrics, 280 posts/week

Business Analytics

AI reports, WACC, competitor segmentation

E-commerce

Ozon: pricing, demand analysis, 1743 SKUs

Career Automation

HH Hunter: scoring → auto-apply pipeline

Monitoring

Grafana + Pulse + Uptime Kuma, 4 servers

News Aggregation

20 sources, HDBSCAN clustering, bias-check

Personal OS

COO: tasks, goals, path auditor, 688 tests

Ready to discuss?

If you need an architect who builds autonomous AI systems — reach out.

Serbia-based · CET/CEST timezone · EU-aligned working hours · International contracts experience