Gostik

Platform for automatic formatting and validation of student papers per GOST 7.32-2017 with university guidelines support. The product is deployed at UrFU.

Product features

• Document auto-formatting – brings Word files in line with GOST 7.32-2017 and university guidelines
• Formatting validation – finds issues in structure, styles, tables, figures, lists, and title pages
• DOCX and PDF comments – shows findings in the document and on the PDF page, anchored to the issue location
• AI-assisted formatting – helps fix document fragments and detect paper metadata

What I worked on

• Built the NestJS BFF for the SaaS platform: JWT/OAuth, document lifecycle, token-based monetization, YooKassa integration (webhook, fiscal receipt), S3 uploads, RabbitMQ job enqueueing, and an idempotent callback contract with the worker
• Designed an async Python document processing worker: step-based pipelines for MD and DOCX, 180+ algorithmic GOST checks, PyMuPDF layout analysis, a hybrid Word pipeline with OOXML comments. 1000+ documents processed
• Built the product frontend, dev interface, and PDF overlay for findings visualization
• Set up analytics and observability: ClickHouse, Vector, Grafana
• Integrated LLM scenarios for document metadata (type, university, confidentiality) and AI-assisted formatting