Senior System Analyst / Systems & AI

Mansur Kakushkin

I design complex systems, integrations and AI-powered workflows — from business problem to architecture and implementation.

View ExperienceSelected ProjectsDownload CV
Scroll to explore

About

I design enterprise systems across the full cycle — from business problem to architecture, APIs, data, integrations and AI automation.

I'm a Senior Systems Analyst with 7+ years of experience in enterprise technology. I have a strong background in designing enterprise systems across the full cycle - from identifying business problems to designing architecture, APIs, data structures, integrations, and AI-powered automation.

Currently, I'm leading the development of AI-powered systems, applying LLMs, RAG, Graph RAG, knowledge graphs, and AI agents to engineering workflows. I'm responsible for translating business logic into system requirements, designing system architecture, defining APIs and data structures, building integrations, and following solutions through to production.

In my previous roles at X5 and Sber, I successfully designed complex information systems, integrations, and automation solutions at scale. I was in charge of the full development cycle: from business processes and requirements to system design, implementation, and final results. One of my key projects was an AI Code Knowledge Platform — an automated code knowledge extraction system that enables engineering teams to understand new codebases in hours instead of days, serving analysts, engineers, support teams, testers and managers. I also contributed to a product that received patent RU2732071C1 for automated legal analysis.

I'm currently looking for Senior Systems Analyst, Solution Architect, or AI-focused engineering roles where I can take responsibility for complex systems, integrations and architecture at scale and contribute to delivering impactful technology solutions.

I work at the intersection of:

System ArchitectureIntegration ArchitectureAPI DesignData PipelinesAI / LLM / RAGKnowledge GraphsMCP / AI AgentsSoftware Engineering

Selected Impact

Resource Optimization

5x

through system automation

Faster Delivery

30–40%

idea-to-feature cycle reduction

Incident Resolution

~3x

faster in production

Patent

RU2732071C1

participant in automated legal analysis system

Experience

Career Path

X5 Tech

Senior System Analyst

Sep 2025 — Present (1 year)
Moscow / Remote
  • Coordinate delivery across 2 project teams and 1 product team — performing project coordination, defining priorities and directing day-to-day work across developers, analysts and QA.
  • Redesigned the requirements-to-development workflow around AI-assisted specifications, Git-based documentation, automated test generation and AI-powered code reviews — reducing idea-to-feature delivery time by 30–40% and enabling the team to handle a workload that would have required ~1.5× the original headcount.
  • Introduced AI-assisted incident investigation with structured technical context and MCP integrations with task trackers, reducing production incident resolution from 2–3 days to under 1 day through faster root-cause analysis.
  • Conceived, designed and built AI Code Knowledge Platform — a full-stack system (Python, React, MCP, Memgraph) that automatically extracts knowledge from codebases, generates documentation and provides AI chat over system context. Scaled from personal workflow → 5 analysts → multiple teams. Now in user testing across analysts, developers, support and QA.
  • Embedded AI across the delivery pipeline — requirements engineering, documentation generation, test case creation, API specification review and automated code analysis — establishing AI as a core tool rather than an isolated experiment.
  • Brought all project documentation to a unified structured format with templates, references and usage rules, enabling consistent onboarding and cross-team handoffs.
  • Scaled AI practices across teams through workshops, training and cross-team collaboration, expanding adoption beyond individual projects.

System Analyst

Oct 2023 — Aug 2025 (1 year 11 months)
Moscow / Remote
  • Designed architecture for a large-scale tax monitoring platform integrating with ~7 external systems (SAP, FNS, signing/EC services), coordinating integration contracts across business processes and internal services. Solution serves 9 legal entities, 4 internal divisions and 40+ users.
  • Designed a schema-driven approach: XSD → JSON Schema → UI generation pipeline, eliminating developer involvement in supporting new schema versions — screen forms now auto-adapt when analysts update the schema.
  • Integrated new data sources for automated report generation, reducing manual workload equivalent to approximately 2 FTE.
  • Defined API contracts and integration interfaces between SAP, PostgreSQL, ClickHouse and internal services, coordinating requirements across teams.
  • Built monitoring dashboards in Grafana and implemented alert logic for production system health tracking.
  • Performed system and business analyst functions: requirements gathering, task creation for developers, prioritization and implementation control.
SAPPostgreSQLClickHouseREST APIAirflowPythonReactMemgraphMCPGraph RAGDocker
Extrum

Business / System Analyst

Nov 2022 — Sep 2023 (11 months)
Astana
  • Collected requirements, wrote technical specifications and coordinated architecture decisions with customers and development teams.
  • Managed backlog, priorities and timelines; assembled the development team and directed deliverables.
  • Tested functionality, identified issues and drove resolution through structured task management.
ClickUpSCRUMDraw.ioFigma
Smart Automation

System Analyst

Nov 2022 — Apr 2023 (6 months)
Remote
  • Configured and implemented Smart Automation product for 3 enterprise customers, creating project documentation and coordinating delivery.
  • Identified 15 critical blocking defects, prioritized and resolved them; added features that improved workflows for users and administrators.
Smart Automation
Sber

System Analyst

Dec 2018 — Oct 2022 (3 years 11 months)
Moscow
  • Co-created an ML-based document generation service achieving 90% F-measure in collaboration with the data science team.
  • Developed and maintained LegalTech projects automating due diligence processes in corporate lending.
  • Built an internal analytics service for the audit team; developed Python-based DSL tooling for automated due diligence.
PythonDSLMLLegalTech

Selected Projects

Case Studies

AI Code Knowledge Platform
Graph RAG / AI Agents / MCP / Knowledge Graph

Designed and built an end-to-end platform for automated code knowledge extraction from source code. Transforms repositories into knowledge graphs, generates documentation through AI agents and provides AI chat with multi-project context and MCP integrations.

Full-stack platform4 doc pipelinesAI chat + MCP
LLMAI AgentsRAGKnowledge Graph+4
View Case Study
AI Delivery Transformation
Skill-packs / Context Structuring / AI Harness

End-to-end AI transformation of delivery processes at X5 Tech. Created standardized skill-packs for AI tools, restructured task decomposition workflows, and built AI Harness infrastructure — reducing idea-to-feature cycle by 30-40%.

30-40% faster delivery3x faster incident resolutionScaled to multiple teams
LLMAI HarnessSkill-packsMCP+2
View Case Study
Tax Monitoring System
SAP / FNS / Microservices / PostgreSQL / ClickHouse / Airflow

System integrated with SAP and corporate systems for Federal Tax Service monitoring. Microservice architecture serving 40+ users across 9 legal entities.

5x resource optimization9 legal entities40+ users
SAPPostgreSQLClickHouseREST API+3
View Case Study
Schema-driven Automation
XSD → JSON Schema → Python

Automated pipeline for document schema processing, transforming XSD schemas into JSON Schema and generating processing scripts — reducing manual analyst work.

5x faster processingReduced manual work
XSDJSON SchemaPythonAutomation
View Case Study

How I Think

From Problem to Result

I don't just document requirements — I work across the full spectrum from business problem understanding to architecture design and implementation.

Business Problem
Business Process
Requirements
System Context
Architecture
API / Data
Implementation
Automation / AI
Result

AI Engineering & Research

System Analysis + Architecture + AI

I research AI technologies and build production-oriented AI systems with them. My applied work includes designing and implementing AI Code Knowledge Platform — a full-stack platform for automated code knowledge extraction, graph-based analysis and documentation generation.

Built
AI Code Knowledge Platform

Designed and built a platform that automatically extracts knowledge from codebases and generates documentation. Used by analysts, engineers, support teams, testers and managers for system understanding and onboarding.

Knowledge ExtractionAutomated DocumentationAI ChatOnboarding
Built
Analyst Skill-packs & AI Agents

Created standardized skill-packs — sets of prompts, context templates and rules for using AI tools in analytics workflows. Built agent configurations for requirements analysis, code review and documentation generation.

Skill-packsAI AgentsContext TemplatesAnalytics Workflows
Built
MCP Servers & Integrations

Built and operate MCP (Model Context Protocol) servers for AI agent integration with enterprise systems — enabling agents to access code, documentation, project management tools and communication platforms.

MCPTool IntegrationAI AgentsExternal Systems
Research
LLM & RAG Systems

Applying large language models with retrieval-augmented generation for code understanding, documentation and system analysis workflows.

LLMRAGVector SearchEmbeddings
Experiment
Graph RAG & Code Graph

Exploring graph-based models for code analysis — transforming codebases into knowledge graphs for dependency analysis, context search and architecture understanding.

Graph RAGKnowledge GraphASTDependencies
Research
AI-assisted Analysis

Using AI to accelerate the analyst workflow itself — from requirement gathering and process modeling to architecture documentation and test generation.

AutomationWorkflowSystem Analysis