<- Featured work
FEATURED / PRODUCT SYSTEM

The Loop

Campus event platform for discovery, RSVP, friends, chat, maps, and carpool workflows.

ReactFastAPIPostgreSQLSQLAlchemyJWT
Live deployment GitHub repo
The system

Connect event discovery, interest context, RSVP, social coordination, chat, and carpool planning.

Product engineering proof alongside the AI systems work.

Visual proof

See the system before the explanation.

Reviewed project artifacts, shown with their existing evidence status.

The Loop events list showing dated campus events and category tags
Live deployment workflow screenshot Primary public visual showing the dated event timeline and category tags.
The Loop event detail page showing event actions and location map
Owner-captured deployment screenshot Supports event detail, location, and action-flow evidence.
The Loop map view showing event pins near Thapar Institute
Owner-captured deployment screenshot Supports map-view event discovery evidence.
The Loop landing/login screen
Owner-captured deployment screenshot Optional support visual for the landing/login state.
The Loop profile page with personal contact details redacted
Prototype workflow evidence Personal contact details are masked; supporting profile-flow evidence only.
The Loop friends page with sent-request username redacted
Prototype workflow evidence Sent-request username is masked; supporting social-flow evidence only.
The Loop system diagram showing React, FastAPI, database, auth, realtime, maps, and carpool layers
System diagram Repo-backed topology diagram for app boundaries.
The Loop recommendation flow diagram showing interests, event tags, cosine similarity, joins, and updated user interests
Recommendation flow Shows interest/tag matching flow; not recommendation-system research proof.
The Loop carpool and chat flow showing join, carpool request, friend request, chat history, and WebSocket delivery
Carpool and chat flow Shows persisted coordination flow and active WebSocket path.
The Loop deployment diagram showing React Vite frontend, FastAPI Uvicorn backend, database, Google OAuth, Maps, and Render services
Deployment diagram Render service topology only; not monitoring, backup, or SLA proof.
01 / Build

What exists

What I built

  • Central event feed with list/map views.
  • RSVP/join flows and event detail pages.
  • Friend requests, chat, and carpool coordination.
  • Admin event workflows.
  • Recommendation flow using interest/tag matching.

Architecture

  • React/Vite frontend.
  • FastAPI/Uvicorn backend.
  • SQLAlchemy data models.
  • WebSockets for chat.
  • Google Maps/OAuth integrations.
  • SQLite/PostgreSQL depending environment.
02 / Ownership

Collaborative Full-Stack Contributor

Atishay Jain collaborated across product flow and implementation on a campus event-discovery and coordination platform; ownership is not cleanly separable by module.

Contribution

  • Worked across the product flow for centralized campus and outside-campus event discovery.
  • Contributed across RSVP/social coordination, carpool, maps, recommendations, and admin/event workflows.
  • Kept the case study focused on workflow evidence instead of unsupported solo module ownership.

Collaborative scope

  • Event discovery, RSVP/social coordination, friends, chat, carpool, maps, recommendations, and admin/event workflows are collaborative platform areas.
  • Exact ownership for auth, RSVP, chat, carpool, maps, recommendations, admin, or deployment is not separated in public evidence.
03 / Decisions

Engineering judgment

Key decisions

  • Frame the project around campus coordination outcomes.
  • Persist event joins, friend requests, carpool groups, and chat history.
  • Keep deployment, screenshot, and realtime limitations visible without production-ops claims.

Proof highlights

  • Deployment and GitHub repo are Verified.
  • Live screenshots cover events list, event detail, map view, landing/login, and redacted profile/friends states.
  • System, recommendation, carpool/chat, and deployment diagrams are repo/docs-backed.
  • Workflow screenshots support prototype evidence, not production reliability, active-user, or SLA claims.
Product workflow

From discovery to coordination.

Discovery

Student logs in, creates profile/interest context, and reaches event discovery.

Interest matching

Interests and event tags shape feed ordering through cosine similarity.

RSVP / join

Joining an event writes a user-event relationship and updates interest weights.

Friends, chat, and carpool

Friend requests, chat history, active WebSocket delivery, and carpool requests connect event intent to coordination.

Admin / event management

Admin endpoints support event create, update, delete, and admin-owned event listing.

Operational boundaries

Prototype/MVP evidence only; no reliability, SLA, or user-scale claims.

Boundary

What this evidence does not claim

Collaborative project; exact module split is not public.

  • WebSocket delivery depends on valid JWT state and active connections.
  • Offline delivery and reconnect/session behavior are not proven as hardened production features.
  • Matching is a product workflow, not recommendation-system research.
  • Render/free-tier constraints may affect perceived reliability.