ตำแหน่ง Senior system analyst

บริษัท X5 Tech (x5.ru)

ที่อยู่บริษัท Moscow

เงินเดือน สำหรับสมาชิกเท่านั้น

ลักษณะงานที่ทำ
Project: Assortment Performance Management Module
Internal analytical and operational solution for category managers and the commercial function of a retail chain. The system consolidated sales, margin, inventory, supply, write-off, and promotion metrics, identified persistently underperforming SKUs, generated recommendations, and supported assortment decisions through handover to the existing product-matrix management system. My primary focus was backend calculation logic, recommendation rules, integrations, and the lifecycle of assortment decisions.
Technology stack: REST API, Apache Kafka, PostgreSQL, SQL, Swagger, Postman, DBeaver, Kibana, BPMN, Use Case, UML Sequence, UML State, ERD, C4, Draw.io, PlantUML, Jira, Confluence, JSON, YAML
Key responsibilities and achievements:
• Designed the target architecture of the module using C4: defined boundaries for user management, SKU data acquisition and calculation, recommendations, SKU analog selection, forecasting, assortment decisions, notifications, and integration with the external product-matrix system. Documented synchronous REST calls, asynchronous Kafka events, and dependencies on enterprise databases. Together with the architect, separated operational data from heavy sales history to avoid duplicating large analytical datasets in PostgreSQL.
• Formalized the SKU performance assessment process: documented the BPMN and Use Case from defining the analysis period to creating recommendations, taking into account promotions, product age, region, and store format, including alternative scenarios for new and temporarily underperforming items. Prepared UML Sequence diagrams for interactions between the recommendation service, forecasting service, analog-selection service, their databases, and Redis caches, reducing ambiguous interpretations of requirements between development and QA by approximately 25%.
• Designed the REST API for the assortment decision card in Swagger: endpoints for retrieving SKU and comparable-product metrics, creating a decision, adding comments, changing status, and confirming the final action. Documented request/response models, required fields, allowed state transitions, error codes, and data mapping from the calculation backend; validated contracts using Postman.
• Worked out the estimated impact of delisting or restricting an SKU: documented the algorithm for considering current revenue, margin, inventory, open deliveries, and substitute products, as well as Use Cases for alternative scenarios and the REST API response structure. A unified calculation scenario reduced the number of manual summary reports prepared by commercial analysts before item review by approximately 30-40%.
• Designed the ERD for the operational model covering assortment decisions, calculation results, applied business rules, recommendations, approvals, and state history. Documented relationships, constraints, result versioning, and rules for storing references to source datasets; prepared SQL checks in DBeaver to validate metrics and status transitions before acceptance.
• Designed event-driven integration via Kafka for changes to product characteristics, assortment-decision statuses, and the result of product-matrix updates. Defined business events, message keys, producer/consumer interaction, JSON structure, handling of delayed events, and consumer error scenarios. Documented the asynchronous UML Sequence flow and monitoring requirements in Kibana, reducing manual reconciliation of decision statuses with the adjacent system from several hours to approximately 20-30 minutes.
• Designed the handover of an approved assortment change to the existing assortment-management domain: documented the API contract and mapping for SKU, store groups, change type, effective date, reason, and substitute product, as well as a UML State model for the states 'approved', 'sent', 'applied', and 'rejected'. Added retry rules for technical failures without requiring repeated business approval, reducing manual repeat operations by approximately 30%.
• Participated in analysis of defects and production incidents involving discrepancies in calculated metrics and decision states between analytical and product-matrix domains. Correlated requests using Postman, logs and operation IDs in Kibana, PostgreSQL data in DBeaver, and Kafka events. Refined contracts, calculation-versioning rules, and technical-audit requirements, reducing diagnosis time for recurring issues from 1-2 hours to approximately 20-30 minutes.
• Defined role and access requirements for category managers and commercial analysts: specified permitted actions by card state, auditing of manual overrides of automated recommendations, and restrictions on modifying completed decisions. Documented requirements in Confluence and Jira and aligned them with business process owners and the development team.

ตำแหน่ง System analyst / business analyst

บริษัท Domclick (domclick.ru)

ที่อยู่บริษัท Moscow

เงินเดือน สำหรับสมาชิกเท่านั้น

ลักษณะงานที่ทำ
Project: Preliminary Collateral Property Valuation Service
Calculation and integration backend domain of a mortgage platform that received property characteristics, normalized and enriched data, selected market comparables, calculated an indicative value, and determined whether manual review was required. The valuation result was passed to mortgage and risk domains and used as one of the inputs for the subsequent credit process.
Technology stack: REST API, Apache Kafka, PostgreSQL, SQL, Swagger, Postman, DBeaver, Kibana, JWT, BPMN, Use Case, UML Sequence, UML State, ERD, C4, PlantUML, Jira, Confluence, JSON, YAML
Key responsibilities and achievements:
• Participated in decomposing the valuation service using C4: separated request management, property data, address enrichment, comparable selection, valuation calculation, and manual review domains. Documented REST and Kafka interactions, state ownership, and external-source failure points, enabling long-running enrichment and calculation operations to be separated from the synchronous mortgage flow.
• Worked out the end-to-end BPMN valuation process from mortgage application intake to delivery of the final value. Documented roles, Use Cases, UML State transitions for the request, and alternative scenarios for insufficient data, ambiguous addresses, low confidence levels, and manual review. More detailed transition logic reduced development and QA rework caused by returned specifications by approximately 20-25%.
• Designed REST APIs for creating and retrieving valuations in Swagger: input/output models for property, transaction price, and property characteristics; required fields; business validation of area and property type; error codes; current-state and final-result retrieval. Validated contracts and negative scenarios using Postman and documented repeated valuation requests after changes to property parameters.
• Designed integrations with external platforms for property-data enrichment: documented the UML Sequence for address normalization, mapping of the source address to a standardized representation, building ID and coordinates, and handling of ambiguous results, timeouts, and technical errors. Added rules distinguishing technical source unavailability from a business scenario requiring address clarification, reducing incorrect routing to manual valuation.
• Detailed the comparable-property selection and market-sample filtering algorithm: formalized criteria for location, area, number of rooms, building type, and data freshness; rules for excluding duplicates and anomalous values; and the input/output model of the calculation service. Documented UML Sequence flows and acceptance criteria for boundary scenarios, reducing discrepancies between analytical requirements and implementation during sample validation by approximately 20%.
• Designed the ERD for valuation requests, properties, normalized addresses, comparable sets, calculation-model versions, calculation results, manual decisions, and state history. Worked out valuation versioning and immutability of previous results; used SQL in DBeaver to validate consistency between property versions, transaction price, and calculation results.
• Designed the Kafka event-driven flow for long-running enrichment and valuation-state changes: documented events for completion of address verification, receipt of additional characteristics, calculation completion, and routing to manual review, including JSON structure and retry rules. Separating the stages prevented the mortgage process from waiting synchronously for all external calls and enabled diagnosis of stuck requests by states and events.
• Designed delivery of valuation results to mortgage and risk systems: documented the API contract for final value, range, confidence level, manual-review flag, and restriction reason; UML Sequence diagrams for successful and failure scenarios; and retry rules for technically undelivered results. This reduced manual clarification between mortgage and valuation domains by approximately 30%.
• Investigated test and production issues involving incomplete valuations and mismatched property versions: checked correlation IDs in Kibana, calculation state and versions using SQL in DBeaver, REST requests in Postman, and Kafka event sequences. Refined the state model, revaluation requirements, and logging requirements for source-data versions, reducing diagnosis time for typical cases by approximately 1.5-2x.
• Together with the process owner and development team, defined audit requirements for manual specialist decisions: stored the original automated result, modified comparable set, manual value, decision reason, and user. Documented constraints and acceptance criteria in Jira and Confluence so that a valuation could be reproduced for a specific version of the input data and calculation algorithm.

ตำแหน่ง System analyst

บริษัท Sibur (sibur.ru)

ที่อยู่บริษัท Moscow

เงินเดือน สำหรับสมาชิกเท่านั้น

ลักษณะงานที่ทำ
Project: Industrial Equipment Condition Monitoring System
Industrial monitoring domain for pumps, compressors, electric motors, and other equipment based on IIoT sensor data. The system ingested telemetry, controlled data quality, detected persistent deviations, generated diagnostic events, and passed confirmed issues to the existing maintenance and repair management domain.
Technology stack: REST API, Apache Kafka, PostgreSQL, SQL, Swagger, Postman, DBeaver, Kibana, Grafana, BPMN, Use Case, UML Sequence, UML State, ERD, Draw.io, PlantUML, Jira, Confluence, JSON
Key responsibilities and achievements:
• Formalized the technical-deviation processing flow from telemetry ingestion to creation of a maintenance task: documented BPMN, Use Cases, operator and quality-engineer roles, UML State for diagnostic events, and alternative scenarios for sensor failures, short-term spikes, and confirmed equipment defects.
• Designed the Kafka telemetry-ingestion model: defined message structure with equipment and sensor IDs, measurement type, timestamp, value, and quality flag; partitioning rules; and handling of out-of-order events. Documented UML Sequence for measurement ingestion and normalization and monitoring requirements for consumer latency and errors in Grafana and Kibana.
• Designed data-quality control rules separately from equipment diagnostics: formalized checks for missing data, delays, physically impossible values, long-unchanged signals, and deviations in ingestion frequency. Separating telemetry-quality events from technical deviations reduced false diagnostic cards caused by sensor issues by approximately 25-30%.
• Documented REST APIs for diagnostic events and integration with the maintenance and repair management system in Swagger: equipment card and event-history retrieval, diagnostic confirmation or rejection, technical-task creation, and receipt of completed-work results. Defined input/output models, statuses, error codes, and cross-system ID mapping; validated contracts in Postman.
• Designed the operational ERD covering equipment, measurement points, sensors, diagnostic rules, events, criticality, specialist decisions, and related maintenance tasks. Documented relationships and constraints and used SQL in DBeaver to validate event-to-equipment linkage and processing history before acceptance.
• Participated in incident analysis involving telemetry delays and reappearance of diagnostic events after repairs: correlated Grafana charts, Kibana logs, Kafka events, and PostgreSQL data. Refined requirements for measurement and message timestamps, event reopening, and normalization checks after completion of maintenance work.

ประสบการณ์ทำงานอื่นๆ
Eduson academy.
Educational course "System and business analysis", 2021

Eduson academy
Educational course "AI tools for analitics", 2023

สถาบัน Eduson academy

หลักสูตร System analysis / business analysis

สถาบัน Eduson academy

หลักสูตร AI tools for analytics

สถาบัน Eduson academy

หลักสูตร Financial analysis and financial modeling