Smart Scrum Master

Inspiration

Modern Agile teams spend hours in sprint meetings, yet critical decisions often get buried in long transcripts and scattered notes. Writing Jira tickets after every discussion is repetitive, manual, and error-prone.

We built Smart Scrum Master to eliminate that friction — transforming conversations directly into structured, actionable sprint workflows.

What it does

Smart Scrum Master is an AI-powered assistant that converts meeting discussions into ready-to-use Jira tickets and searchable knowledge. -Transcribes and semantically indexes sprint discussions. -Generates structured Jira tickets with descriptions and acceptance criteria. -Enables natural language search across past meetings. -Updates and manages tickets through a conversational AI interface. -Supports offline ticket creation with automatic sync.

It turns unstructured conversation into execution-ready output.

How we built it

We built a full-stack AI system integrating semantic search, agent reasoning, and workflow automation. -FastAPI (Python) backend for APIs and orchestration. -Angular + TypeScript frontend for interactive UI.OpenSearch for semantic indexing and retrieval. -Gemini APIs for LLM reasoning and embeddings. -Jira Cloud APIs for ticket creation and updates. -IndexedDB for offline-first ticket storage.

Semantic similarity between transcript chunks and queries is computed using cosine similarity: $$\text{similarity}(A, B) = \frac{A \cdot B}{|A| |B|}$$ We also implemented a ReAct-style agent that dynamically selects tools to create, update, and retrieve tickets based on user intent.

Challenges we ran into

-Building a production-style AI workflow within 24 hours. -Designing reliable offline synchronization and conflict handling. -Ensuring accurate ticket generation from noisy meeting transcripts. -Coordinating agent tool selection and structured API calls. -Time constraints forced rapid architectural decisions and tight team collaboration.

Accomplishments that we're proud of

-Delivering a complete AI-driven productivity tool in 24 hours. -Automating ticket creation directly from sprint discussions. -Implementing semantic search over meeting history. -Designing an offline-first architecture with sync tracking. -Building an intelligent agent capable of multi-step reasoning.

What we learned

-AI systems require orchestration, not just model calls. -Developer productivity tools demand strong UX design. -Offline-first design significantly improves reliability. -Agent-based architectures introduce new debugging challenges.

What's next for Smart Scrum Master

Our goal is to evolve Smart Scrum Master into a fully autonomous AI teammate. Next steps include: -Multi-platform integration: GitHub Issues, Asana, and Trello. -Contextual Awareness: Improve ticket quality using team-specific history. -Speaker Attribution: Automatic action-item tagging per person. -Platform Integration: Direct hooks into Zoom, Google Meet, and Slack. -Analytics: Providing sprint velocity and productivity insights.