
FORCE EventHub - Building a Smarter Way to Plan and Coordinate Events in Cameroon
How Team Gamma built an AI-powered event and logistics platform to connect organizers and service providers during the AWS re:Deploy Hackathon 2026 using Kiro and modern cloud technologies to optimize the Planning and Coordination of Events in Cameroon.
EventHub : Building a Smarter Way to Plan and Coordinate Events in Cameroon
How Team Gamma built an AI-powered event and logistics platform to connect organizers and service providers during the AWS re:Deploy Hackathon 2026 using Kiro and modern cloud technologies to optimize the Planning and Coordination of Events in Cameroon.
Smarter Planning. Better Decisions. Good Connections
By Team Gamma

"Every great Event has to be well planned. We're building the technology that helps orgnanise them faster, more fairly, and with the best providers."
1. The Problem Behind Event Planning in Cameroon
Planning an event is rarely about a single service. Whether it is a wedding, a corporate gathering, a ceremony, or another organized event, the organizer may need to coordinate several independent actors: caterers, decorators, photographers, security teams, transportation providers, venues, and other service providers.
The difficulty is not only finding these services. It is coordinating them around the same event, date, location, budget, and schedule.
In a fragmented workflow, an organizer may have to contact providers individually, compare their offers and availability, keep track of confirmations, coordinate logistics, and communicate changes to different people. As the event gets closer, even a simple change in venue, schedule, or availability can affect several participants at once.
This creates a coordination problem involving three main groups of users. Organizers need to plan and manage the services required for an event. Providers need to discover missions, coordinate with other providers when necessary, and receive relevant information. Guests need reliable information about the event location, routes, and last-minute changes.
For the Month 2 challenge, we decided to approach this problem from a single perspective: instead of treating event planning, provider discovery, and event logistics as separate tasks, what if they could be connected through one platform?
This question became the starting point for EventHub.
2. Introducing EventHub
EventHub is a centralized event and logistics platform designed to bring the different parts of event coordination into one connected workflow.
Instead of managing service providers, event logistics, and attendee information separately, an organizer can create an event, define the services it requires, discover suitable providers, and track their status from a single platform.
EventHub connects two primary participants.
Organizers use the platform to plan their events, manage required services, select providers, and coordinate logistics. Providers can present their services, receive event requests, collaborate with other providers, and manage their missions.
The platform also introduces an AI-powered recommendation layer. Rather than presenting organizers with a simple list of providers, EventHub is designed to progressively consider factors such as service requirements, availability, budget, distance, ratings, and reliability when identifying suitable candidates
3. From the Problem to the Product
Once the coordination problem was identified, we deliberately limited the MVP to the workflows that could demonstrate the highest practical value.
The first decision was to centralize event services. Instead of managing caterers, decorators, transport providers, security teams, and other vendors independently, an organizer can manage the services associated with an event from a single workspace.
The second decision concerned provider discovery. EventHub was not designed as a simple vendor directory. Provider selection needs to consider more than service category, including factors such as availability, price, distance, ratings, and reliability. This led us to introduce a recommendation pipeline rather than a conventional search-only experience.
The third decision was to connect planning with logistics. Once providers and guests are associated with an event, location and last-minute communication become part of the same workflow. This is why EventHub includes venue mapping, routing, real-time notifications, and WhatsApp notifications via EventHub's whatsApp Bot


Finally, we wanted recommendations to improve from actual organizer behavior. Actions such as accepting, removing, or replacing a suggested provider can become feedback for future ranking.
This resulted in an MVP focused on four connected capabilities: event and service centralization, provider discovery, logistics support, and intelligent recommendations.
The product specification defines these workflows through use cases covering event services, provider matching, routing, provider teams, ratings, last-minute notifications, WhatsApp communication, and AI-assisted recommendations.

4. How EventHub Works
EventHub follows the lifecycle of an event across three main participants: the organizer, the provider, and the guest.
The organizer starts by creating an event and defining the services required. Providers can then be discovered for each service and contacted for a specific mission. Their status can be tracked as the event progresses.
Provider selection is supported by a matching workflow. Instead of returning an undifferentiated list, EventHub can evaluate providers using factors such as service fit, rating history, availability, price, and proximity. The organizer remains responsible for the final decision.
Providers have a separate workflow. They can manage their service profile, receive missions, accept assignments, and collaborate with other providers through provider teams when several services need to be coordinated.
Guests and providers also have access to event location and routing information. For providers handling several locations, the system is designed to support multiple waypoints.
The final part of the workflow is communication. When an event changes at the last minute, relevant participants need to be informed quickly. EventHub therefore combines in-app notifications with WhatsApp communication for critical updates. These workflows are reflected in the product specification through the event, provider, routing, notification, and recommendation use cases.


5. The AI-Powered Recommendation Layer
The recommendation layer is one of the parts of EventHub that we designed to evolve rather than attempting to build a complex machine-learning system immediately.
A simple similarity search would not be sufficient for our use case. An organizer may prefer a provider because of a combination of price, rating, distance, availability, and reliability. A recommendation system also needs to learn from what organizers actually choose.
We therefore designed the recommendation architecture as a six-layer pipeline.
The first layer is a conversational agent that collects information about the event, such as its type, date, budget, guest count, and style.
The second layer is the Event Blueprint Engine. It translates the event description into service categories. For example, a wedding may require catering, decoration, photography, entertainment, and transportation.
The third layer retrieves candidates using hard constraints such as location, availability, and budget.
The fourth layer ranks these candidates using multiple signals rather than relying on a single similarity score.
The fifth layer records organizer actions. Accepting, removing, or replacing a recommendation creates feedback that can later improve the ranking process.
The sixth layer provides orchestration. The conversational agent exposes the recommendation pipeline through tools such as
get_event_blueprint() , search_providers() , rank_candidates() , and record_feedback() .Importantly, the architecture separates what can be demonstrated in the MVP from what requires more data. The specification proposes starting with a rule-based blueprint and weighted scoring, then introducing learning-to-rank and eventually exploration mechanisms as sufficient interaction data becomes available.
This distinction was important to us: we did not want to claim a trained recommendation model where the available data could not justify one.


6. Technical Architecture
EventHub was implemented as a multi-layer application connecting the user interfaces, backend services, persistent data, real-time communication, external integrations, and recommendation workflows.
At the application level, the architecture separates the frontend from the backend and from external integrations. The product specification defines the core application around a client layer, API layer, data layer, integration layer, real-time layer, and AI recommendation layer.
For the deployed infrastructure, the team used AWS services for identity, storage, backend infrastructure, messaging, scheduling, and observability. The resulting flow can be summarized as:
User → Frontend → Authentication/API → Backend Services → Data & Integrations Amazon Cognito handles user identity and authentication. Amazon RDS provides relational data storage, while Amazon S3 is used for object storage such as images and other application files. CloudWatch provides monitoring and observability.
For asynchronous and longer-running workflows, the architecture also incorporates messaging, scheduling, and Temporal-based orchestration where required.
The application connects external capabilities such as WhatsApp, maps/routing, and AI services to the core event-management workflow.
The important architectural decision was to keep these integrations separated from the core business logic. This allowed us to change an external service without redesigning the entire application.

7. Building EventHub with Kiro
Kiro was used as part of the development process rather than simply as a code-generation tool.
We used the specification-driven workflow to move from product requirements toward implementation, then iterated as technical constraints appeared during development. The process involved exploring requirements, structuring implementation tasks, generating and modifying code, testing integrations, and debugging unexpected behavior.
This was particularly useful for a project with several interconnected components. EventHub combines event management, provider matching, authentication, maps, real-time communication, WhatsApp, and AI workflows. Changes in one part of the system could therefore affect several others.
Kiro helped us maintain a more structured development loop: define what needed to be implemented, generate or modify the corresponding implementation, test it, identify failures, and iterate.
However, the most important lesson was that AI-assisted development does not remove the need for engineering judgment. Generated code still had to be validated against the actual infrastructure, APIs, credentials, runtime behavior, and integration constraints.



8 AWS & Integrations
AWS played a central role in the infrastructure used for the project.
Amazon Cognito was used for authentication and identity management, allowing the application to handle user registration, authentication, verification, and access tokens without implementing password management directly in the application backend.
Amazon RDS was used for structured relational data, while Amazon S3 provided object storage for application files such as images and documents.
Amazon CloudWatch was used for infrastructure monitoring and logs. The architecture also uses AWS-based messaging and scheduling capabilities for asynchronous processing and time-dependent workflows.
The deployment workflow was designed around version-controlled development, build, deployment, and monitoring. The intended development lifecycle was:
Develop → Test → Build → Deploy → Monitor → Improve GitHub was used for source control, while the AWS infrastructure provided the cloud services required by the application.
9. What Worked, What Didn't, and What We Learned
EventHub also depended on external integrations, most notably WhatsApp, mapping/routing services, and AI services. These integrations were not always straightforward, and some of the most significant engineering lessons came from getting them to work reliably.
WhatsApp: From Configuration Confusion to a Working Integration
WhatsApp was one of the most difficult integrations during development.
The main problem was not the application code itself, but the distinction between Meta's testing environment and its production/business configuration. The interfaces exposed different concepts for test numbers, business numbers, phone number IDs, temporary tokens, system-user tokens, and WhatsApp Business assets.
At one point, the development configuration mixed values from the test environment with values required by the production configuration. This resulted in a bot that appeared correctly configured but did not respond as expected.
The permanent access-token workflow also required moving from the developer interface to Meta Business settings and configuring the appropriate System User permissions. Because Meta does not expose all of these settings in the same interface, identifying the correct configuration path took significant time.
The phone-number registration process introduced another source of confusion. A number could already exist in WhatsApp Manager while the developer interface still appeared to request registration, creating ambiguity between adding, verifying, and registering the business number.
Even after the WhatsApp service became reachable, webhook validation produced another issue. The backend returned:
WhatsApp webhook invalid signature The root cause was an incorrect
WHATSAPP_APP_SECRET value on the server. The environment variable existed, but its value was not the actual Meta App Secret. This was particularly difficult to diagnose because the configuration looked correct at first glance.The main lesson was straightforward:
third-party integrations must be treated as separate infrastructure systems with their own environments, credentials, identifiers, and validation mechanisms. Having a variable with the correct name is not evidence that its value is correct.
Maps: Changing the Approach
Maps and routing provided another practical lesson.
Google Maps worked well during the Django and React Native integration stage. However, when we moved to the web JavaScript implementation, the required Google Maps configuration introduced a cost barrier for our hackathon context.
Rather than allowing the mapping service to become a blocker, we changed direction and adopted Leaflet for the web mapping experience.
This was an important engineering trade-off: the goal was not to preserve the first technology choice at all costs, but to preserve the required product capability while keeping the MVP practical.
The biggest lesson was therefore not simply that cloud services or AI tools can accelerate development. It was that
successful product engineering depends on continuously validating assumptions against the real behavior of the systems being integrated.
10. What's Next for EventHub
The hackathon MVP establishes the core workflow connecting organizers, providers, and event logistics. The next phase is about making that workflow more intelligent and reliable rather than simply adding more features.
The recommendation engine is the main area for future evolution. Once sufficient booking and interaction data becomes available, the current ranking approach can evolve toward learning-to-rank and eventually exploration mechanisms for new or less-known providers.
Other future improvements include deeper provider collaboration, richer logistics optimization, improved notification reliability, and broader automation of event planning through the conversational interface.
These capabilities remain future directions unless explicitly demonstrated in the submitted MVP. Keeping this distinction clear allows us to evaluate what EventHub actually achieved during the hackathon while still showing where the architecture can go next.
EventHub started with a simple observation: organizing an event involves many independent actors, but the coordination between them is often fragmented.
We built EventHub to bring those interactions into a single workflow connecting organizers, providers, and guests. The platform combines event-service management, provider discovery, logistics support, real-time communication, WhatsApp notifications, and an AI-oriented recommendation architecture.
The project also demonstrated that building such a platform is not only a matter of designing features. Real-world integrations introduce their own constraints, configuration models, costs, authentication mechanisms, and failure modes.
Working with Kiro and AWS allowed us to move from a product specification toward a working cloud-based prototype while iterating through these constraints.
The result is not the final form of EventHub. It is a foundation: a working direction for making event planning in Cameroon more connected, more structured, and progressively more intelligent.
Our next challenge is to turn that foundation into a system that learns from real events, real providers, and real organizer decisions.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article