Week #3

Week #3 #

Implemented MVP features #

Integrate the frontend with the backend APIs, finalize all datasets, and deploy the pre-trained models.

Demonstration of the working MVP #

iOS #

RegistrationLesson

Android #

Android MVP

ML #

Structured Knowledge Base:

  • Built hierarchical word database (29,997 entries) via Oxford Dictionary parsing
  • Organized by:
    • Topics/subtopics
    • CEFR levels
    • Parts of speech
    • Popularity metrics

Intelligent Processing:

  • Implemented translation pipeline:
    • Fallback from Yandex (CAPTCHA) to alternative API
    • Auto-generated:
      • Accurate translations
      • 5 EN→RU contextual examples per word
      • 3 synthetic variants per sentence (NLP-augmented)

Lightweight Architecture:

  • Metadata-driven recommendations (no heavy ML)
  • Enables:
    • Thematic learning paths
    • Adaptive difficulty scaling
    • Transparent customization

Link to the training code: link

Links to the initial model artifacts: link

Internal demo #

Notes from internal demo

Weekly commitments #

Individual contribution of each participant #

Danila Kochegarov (Team Lead & Backend Developer) #

  • Made Oauth2 with refresh token rotation
  • Reviewed pull requests from iOS, DevOPS and backend developers
  • Made linking telegram account to shared one (with google oauth)
  • Made Telegram bot with lesson learning logic
  • Made setup of async (with many workers) FastAPI backend for ML

Savva Ponomarev (iOS Developer & Product Manager) #

  • Integrated Google authentication
  • Developed login screen UI
  • Built interactive lesson screens:
    • Fill-in-the-blank exercises
    • Translation typing challenges
    • Multiple-choice translation quizzes
  • Configured app navigation/routing
  • Created functional mock implementation
  • Implemente Backend integration

Issues:
https://github.com/FluentlyOrg/Fluently-fork/issues/108
https://github.com/FluentlyOrg/Fluently-fork/issues/109
https://github.com/FluentlyOrg/Fluently-fork/issues/110
https://github.com/FluentlyOrg/Fluently-fork/issues/111
https://github.com/FluentlyOrg/Fluently-fork/issues/113 \

George Selivanov (System Analyst) #

  • Enriched data using Oxford Dictionary (Selenium):
  • 29,997 words with topics, levels, and parts of speech.
  • Failed Yandex Translator (CAPTCHA blocked) → switched to alternative API:
  • Auto-parsed translations + 5 EN/RU example sentences per word.
  • ML integration: Generated 3 synthetic variants per sentence for exercises.
  • Final dataset powers adaptive lessons and recommendations.

Timofey Ivlev (DevOps Engineer) #

Anton Korotkov (ML Engineer) #

  • Hierarchical structure:
    • Topics/subtopics
    • CEFR levels
  • Metadata-driven (no heavy ML):
    • Uses word popularity/complexity
  • Selection flow:
    1. Focused thematic clusters
    2. Broader topics → same level → higher levels
  • Benefits:
    • Transparent & customizable
    • Lightweight
    • Scalable

Daniil Tskhe (Backend Developer) #

  • Helped with setting up authorization via Google
  • Rethinking lesson generation
  • Rewrote the models to meet new requirements
  • Wrote the main JSON generation for the lesson

Evgeniy Bortsov (Android Developer) #

  • Authorization flow
  • Home screen
  • Lesson screens:
    • New Word
    • Choose Translation exercise
    • Fill the gap exercise
  • Backend mock + parsing of backend models into app’s internal models

Issues:
https://github.com/FluentlyOrg/Fluently-fork/issues/108
https://github.com/FluentlyOrg/Fluently-fork/issues/109
https://github.com/FluentlyOrg/Fluently-fork/issues/110
https://github.com/FluentlyOrg/Fluently-fork/issues/111
https://github.com/FluentlyOrg/Fluently-fork/issues/113 \

Plan for Next Week #

Testing #

  • Unit tests: Cover critical backend logic & frontend components
  • Integration tests: API endpoints validation
  • E2E tests: Core user journey (minimum viable coverage)
  • ML: Basic model validation/testing

CI/CD Pipeline #

  • CI Setup: GitHub Actions for auto-build/test on push/PR to main/develop
  • Bonus (+5pts): CD to staging on main merge

Deployment #

  • Staging: Setup on Heroku/AWS free tier
  • Bonus: Production env on VPS
  • Deploy current version to staging
  • Optional: Configure public domain

Team Health #

  • Vibe check:
    • Progress review
    • Blockers discussion
    • Team dynamics alignment

Confirmation of the code’s operability #

We confirm that the code in the main branch:

  • In working condition.
  • Run via docker-compose (or another alternative described in the README.md).