All work

Case study 01

SenseiBI

Ask your spreadsheet a question in plain English. Get a dashboard back.

Nov 2025 – Mar 2026 · Final Year Project · Backend Developer

Backend repo Frontend repo

Walk through it

Manage Dataset screen showing uploads moving through the ingestion pipeline

  1. Upload a CSV or Excel file, ask in plain English, get a dashboard — no SQL, no chart builder.

  2. 8 purpose-built LLM agents: insights, causal analysis, executive summary, normalization, semantic, structure, suggestion, trend.

  3. Extended NL4DV pipeline turns a natural-language question into a real chart.

  4. Multi-tenant from the ground up — businesses, members, invitations, role-based permissions.

  5. Celery + RabbitMQ keep ingestion, dashboard generation and email off the request path.

Built with Shaheer Beig (Agentic AI) and Shaheer Mumtaz (Frontend) .My scope: backend architecture, data pipeline and the agent layer.

Every screen (11)