Conversational Route and Schedule Assistant for Public Transit

Added Jun 2025 3 design docs

Public transit riders mostly have simple questions: when is the next bus, which line reaches the stadium, is this route running today. Yet the answers are buried in PDF schedules and clunky trip planners. Every confused rider becomes a support call or a lost fare, and transit agencies rarely have the staff to answer questions around the clock. The intern builds a conversational assistant for transportation websites using React and TypeScript, with agentic AI at the core. The agent interprets natural-language questions about routes, schedules, and service changes, consults structured route data, and replies conversationally, asking clarifying questions when a request is ambiguous and handling follow-ups in context. The intern designs the chat interface for accessibility and mobile use, and structures the agent's tools so every answer stays grounded in actual schedule data rather than plausible-sounding guesses. The project is a practical introduction to agentic AI product design: defining what the agent can and cannot do, grounding its responses in authoritative data, and building the type-safe, responsive interface that a public-facing service demands.

Related projects

Real-Time Transit Schedule and Route Planning Application

Commuters lose time every day to buses and trains they cannot see coming: schedules change, delays go unannounced, and route planning across lines is guesswork. This project builds a public transport companion that puts live schedule information and route planning in riders' hands while giving transit staff the tools to keep that information accurate. The intern implements a Django backend with PostgreSQL modeling routes, stops, schedules, and user accounts, exposing APIs the frontend consumes for real-time schedule lookups and notifications about delays or changes. The rider-facing interface is built with Next.js, React, and TypeScript, styled with Tailwind CSS, and includes visualizations of routes and timetables plus a route planner that helps users chain services together. An authenticated admin panel lets transit operators publish and adjust schedules, which flow straight through to what riders see. Building it end to end teaches relational schema design for scheduling data, API design for frequently refreshed information, and role-separated interfaces, leaving the intern with a realistic civic-tech product that demonstrates fullstack proficiency.

transportationdjangofullstack+5

View project

Hard

Logistics Route Planning Engine With Traffic and Weather Awareness

For logistics operators, routing is a daily optimization problem with real money attached: traffic, weather, delivery windows, and fuel costs shift constantly, and a plan that was optimal at six in the morning is stale by nine. Dispatchers need routes that adapt, and finance teams need the cost story behind every plan they approve. The intern builds a route planning platform where generative AI produces and revises delivery plans. Given a fleet, a set of stops, and live constraints such as traffic conditions, weather, vehicle capacity, and driver hours, the AI generates route plans with sequencing rationale and cost analyses, then regenerates them as conditions change. A Node.js backend manages plan generation and real-time updates, MongoDB stores fleets, routes, and revision history, and the React frontend presents interactive maps where dispatchers inspect, adjust, and approve routes before exporting driver schedules. Secure user management keeps operator, dispatcher, and driver views appropriately distinct. The project blends optimization thinking with modern AI tooling and teaches the intern to build software where generated plans must survive contact with real-world constraints and human oversight.

energytransportationgenerative-aimongodb+2

View project

Medium

Container Orchestration Pipeline for a Fleet Management Platform

Transportation companies run software that dispatch teams and drivers depend on around the clock, yet many still deploy updates by hand, taking systems offline during business hours and struggling to add capacity when demand spikes. This project tackles that operational gap by automating the full release and scaling lifecycle of a transportation management application. The intern containerizes the application with Docker and deploys it to a Kubernetes cluster configured for horizontal scaling, so the platform can absorb rush-hour traffic without manual intervention. A Jenkins pipeline automates building, testing, and rolling out each release, while Ansible playbooks keep server and environment configuration consistent and repeatable across staging and production. To close the loop, Prometheus scrapes cluster and application metrics and Grafana dashboards surface latency, resource usage, and deployment health, giving operators a live view of the system. By completing the project, the intern demonstrates end-to-end DevOps competence: writing Dockerfiles and Kubernetes manifests, designing CI/CD stages, codifying configuration, and defining the alerts and dashboards production teams rely on every day.

transportationansibledevops+5

View project