Dmitrii Balabanov

📍 Israel | 📧 [email protected] | 📱 +972 54 900 2998 | ✈️ @dimonb

Summary

CTO and software architect with 20+ years of experience building large-scale distributed systems, fintech infrastructure, and high-availability platforms.

Former Yandex architect responsible for billing and payment platforms processing millions of transactions annually with PCI DSS Level 1 compliance.

Currently leading engineering and platform architecture at EBAC Online, where I also drive the shift to AI-native engineering: LLM-powered products in production and spec-driven development with coding agents adopted across the team.

Core Expertise

Distributed Systems | Microservices Architecture | Fintech Infrastructure | Cloud Platforms (AWS, GCP, Cloudflare) | Kafka & Event-Driven Systems | Observability & Reliability Engineering | PCI DSS Level 1 Compliance | LLM Products & Agentic Systems

AI & Agentic Engineering

  • Spec-driven development with coding agents. Every change starts as a written spec and design, agents implement and self-verify it, and humans approve at explicit gates — issue → spec review → implementation → verification → archive.
  • Coding agents, hands-on. Daily production use of Claude Code and OpenAI Codex — custom subagents, skills and slash commands, MCP servers, hooks, and multi-agent fleets running in parallel git worktrees for independent tasks and code review.
  • Autonomous agents. Design and operate long-running, self-looping agents that pick up a backlog, execute, verify, and report without supervision, plus scheduled agents for recurring research and reporting workloads.
  • Agent runtime internals (open source). Contributor to PicoClaw, a lightweight Go agent runtime — merged upstream fixes for the channel delivery contract and secret redaction — and author of an extended fork (~13k lines across 145 files) that runs a production Telegram agent:
    • Deep Telegram integration. Reworked outbound delivery to be synchronous and to stamp channel-native message IDs into conversation history, making the agent a first-class chat participant: it threads replies, works inside forum topics, reacts with emoji, and edits messages it already sent. Replies render as native Markdown through the Bot API Rich Messages endpoint — tables, headings, and quotes survive instead of being flattened.
    • Context engineering. Hierarchical compaction — leaf summaries rolled up under a bounded share of the budget, history budgeted against the real prompt rather than the raw context window, and summarization moved off the turn. Removed 20–35 s of user-visible latency per affected turn and kept the prompt prefix stable so the provider-side cache survives.
    • Also a Codex-protocol WebSocket provider behind a hardened fallback chain, and end-to-end OpenTelemetry tracing across turns, LLM calls, tools, and the message bus.

Experience

EBAC Online 02.2022 - Present

Chief Technology Officer (CTO) Platform responsible for educational delivery and student success, supporting thousands of active learners across a distributed infrastructure.

  • Designed and migrated a company-wide data mesh architecture to AWS, enabling high-traffic educational workloads and scalable data pipelines.
  • Reduced incident resolution time more than twice by implementing full-stack observability (metrics, logs, distributed tracing, alerting) across infrastructure.
  • Led a distributed engineering team of 20+ engineers across multiple time zones, introducing key reliability standards including incident reviews, on-call rotations, and OKR-driven planning.
  • Designed and deployed an AI-powered tutoring platform using LLMs to deliver personalized learning experiences.
  • Architected and shipped a real-time webinar chat on Cloudflare Workers and Durable Objects: an event-sourced, append-only journal with exactly-once idempotent writes and cursor-replicated fan-out shards, load-verified at ~4,000 concurrent participants in a single room with sub-second median delivery.
  • Embedded an AI assistant into the live webinar chat through a dedicated bot API, giving it first-class participation — answering learners in real time, pinning highlights, and applying moderation actions as ordinary journaled events with no side channel.
  • Rolled out spec-driven development with coding agents across the engineering org, keeping architectural decisions explicit and reviewable while agents carry the implementation.

Monetization Platform Department, Yandex, Moscow 01.2021 - 02.2022

Software Architect Enterprise fintech platform responsible for high-availability billing and payment processing across multiple integrators.

  • Architected and developed the internal Yandex Payment Service Provider (PSP) platform processing millions of transactions annually across multiple payment providers.
  • Designed a fault-tolerant, PCI DSS Level 1-compliant infrastructure in Yandex Cloud and completely redesigned the cryptographic schema from scratch to secure sensitive financial data.
  • Built an event-driven microservices platform enabling fully decoupled services using Kafka and Kubernetes.
  • Defined architecture standards adopted across multiple engineering teams as an active member of the Yandex Python Committee (internal language governance group) and the Arcadia (Yandex monorepo) working group.

Automation and Support of Internal Processes Department, Yandex, Moscow 03.2018 - 01.2021

Deputy Head of Department

  • Led the architectural design and launch of the Master Data Hub (MDH), establishing a single “golden record” for enterprise data at scale.
  • Designed the MVP for the highly scalable Yandex.Billing platform, incorporating FaaS and advanced internal cloud infrastructure to process core financial flows.
  • Implemented and standardized rigorous architecture design review processes across multiple engineering teams.
  • Orchestrated the migration of the departmental development workflow to Arcadia (Yandex’s monorepository) for unified CI/CD and dependency management.

Payment, Partner, and Billing Systems Division, Yandex, Moscow 08.2017 - 03.2018

Head of Direction

  • Architected and launched an internal electronic document management platform handling high volumes of sensitive data.
  • Established and streamlined incident management and on-call engineering processes for the division, increasing platform reliability.

Billing Department, Yandex, Moscow 11.2015 - 08.2017

Head of Service Platform responsible for core Yandex billing components, tax compliance, and seamless payment processing.

  • Designed a highly available, distributed, and fault-tolerant architecture for the enterprise billing system using Python, PostgreSQL, and Kubernetes.
  • Launched scalable compliance services for cash registers to meet Federal Law 54-FZ regulations, alongside a high-throughput user API key counter service.
  • Spearheaded the development of core billing microservices serving price/tax management, invoicing, and online payment processing.
  • Scaled the engineering group through active technical recruitment, hiring, and establishing structured performance review cycles for engineering teams.

Billing Department, Yandex, Moscow 07.2012 - 11.2015

Team Lead

  • Architected and developed mission-critical, high-throughput components of the core billing engine in Python.
  • Built and maintained a highly secure, PCI DSS-compliant credit card storage system.
  • Developed and scaled an internal Payment Service Provider (PSP) microservice processing high-volume transaction pipelines.
  • Managed technical recruitment, conducting hundreds of interviews to build a high-performing distributed systems engineering team.

Billing Department, Yandex, Moscow 02.2010 - 07.2012

Senior Software Developer

  • Engineered and optimized scalable core components of the distributed billing engine.
  • Designed and implemented key performance features for the high-volume document flow system.

Billing Department, Yandex, Moscow 03.2008 - 02.2010

Software Developer

  • Developed scalable backend services primarily in Python and XScript.
  • Worked extensively with XML/XSLT and CORBA to build robust legacy distributed systems.

Intersoft Lab, Moscow 2006 - 2008

Team Lead

  • Led the development of sophisticated business reporting tools tailored for Business Intelligence (BI) and OLAP on massive datasets.
  • Engineered a high-performance database access library from scratch to interface Python with Oracle databases, significantly reducing query latency.

Intersoft Lab, Moscow 2002 - 2006

Software Developer

  • Developed robust desktop applications and systems software using C++ and Pascal.
  • Gained deep low-level expertise in Visual Component Library (VCL) internals within enterprise environments.

Education

Moscow Institute of Physics and Technology (State University) (MIPT) 1998 - 2004

Master of Science in Applied Mathematics and Physics

  • Core focus: Economics, Mathematics, and Physics

Skills

  • Languages: Python, Go, C++, TypeScript, JavaScript, SQL
  • Architecture & Leadership: Distributed systems, high-availability platforms, software architecture, technical leadership
  • AI & Agents: Claude Code, OpenAI Codex, MCP, subagent orchestration, autonomous long-running agents, agent runtime internals (Go), context engineering & prompt-cache optimization, LLM application design (RAG, tool use, evaluation), spec-driven development
  • Cloud & Infrastructure: AWS, GCP, Cloudflare Workers & Durable Objects, Kubernetes, Docker, Terraform
  • Databases & Messaging: PostgreSQL, Oracle, MySQL, MongoDB, Kafka
  • Compliance: PCI DSS, SOX

Languages

  • Russian: Native
  • English: Professional working proficiency
  • Hebrew: Beginner (Alphabet+)

Interests

  • Yachting (IYT Skipper), Snowboarding, Photography, Hiking, Lindy Hop dancing