Peter Mamrukov

Enterprise AI & Automation Leader | AI Agents, MCP & Agentic Engineering | ex-CIO/CTO, Hands-On Builder

Why is the human still the glue?

I lead enterprise business applications and build AI agent systems hands-on.

My background includes CIO/CTO roles, enterprise transformation and large technology teams.

I bring enterprise context to AI and hands-on engineering to enterprise transformation.

Local Commander

AI Agents That Can Actually Do the Work

It gives AI agents controlled access to authorized development environments, allowing them to work with files, Git, commands and long-running processes instead of merely telling a person what to do next.

Dyk

Operational Memory for Conversations

Dyk turns authorized messenger conversations into useful operational context: what needs doing, what changed, what matters now, who it relates to and where the information came from.

I am also developing an Engineering Platform that applies the same ideas to the operational context behind an entire product stack.

IDF Eurasia

Financial Services Group

Head of Business Applications

October 2022 - Present

Lead business applications and enterprise automation across finance, treasury, budgeting, IFRS, payments and customer-facing integrations for a financial services group with 10 subsidiaries.

My focus is increasingly on applying AI agents, MCP-based tooling and business process automation to the systems the organization already depends on - reducing manual coordination around development, approvals, integrations and operational workflows.

ELMA365 - building a modern engineering lifecycle around a low-code BPM platform: source snapshots, development workspace, change history, review, Storybook, controlled deployment and target-state verification.

1C:Holding Management - building task-oriented finance workflows around the ERP so approvals, payment requests, IFRS accruals and document history are organized around what a person needs to act on.

EZRun AI

AI Sports Technology

Founder

EZRun AI separates deterministic training evidence from AI interpretation and uses investigation when the available evidence does not justify a confident answer.

A person should not have to explain the same important fact in every conversation. But an AI should not remember everything forever either.

Visit EZRun AI

Airspan

Telecommunications / 5G

Team Leader Software Lab

July 2022 - February 2024

Led a team of four network engineers responsible for infrastructure behind a 5G software-development lab, giving engineering teams the environment they needed to build and test complex telecom software.

BNP Paribas Cardif

Insurance

CIO

June 2021 - April 2022

Led a 35-person technology organization across software development, project delivery, data, infrastructure and cybersecurity.

Connected ERP and CRM systems to enable B2B sales through APIs and reduced IT OPEX/CAPEX by 25% through procurement and license consolidation.

ProbioPharm

Pharmaceuticals

CTO

February 2020 - June 2021

Led post-merger technology integration across multiple businesses and five production sites.

Consolidated three ERP systems, connectivity, support, licensing, procurement and IT operations into a shared operating model.

MedMe Group

Medical Technology

CIO

August 2018 - April 2020

Led technology operations for MedTech businesses and regional technology teams.

Designed the technology foundation connecting operations across three countries through ERP/CRM, medical platforms and contact-center processes.

RSM

Audit & Consulting

IT Project Manager to 2012 · Head of IT to 2014 · CIO to 2018

September 2008 - August 2018

Helped corporate clients turn finance and management processes into working business systems across reporting, budgeting and treasury.

Later led internal IT operations and enterprise application delivery with a team of 30 engineers, combining consulting, software development and infrastructure responsibility.

Scale of responsibility

15+ years of technology leadership · 35-person technology organization · Business applications across 10 subsidiaries · 3 ERP systems · 5 production sites