cover image

Telliscope

  • AI-powered semantic search
  • Personalized user dashboard
  • Multi-source data integration
Expertise
  • Web Application
  • AI
Industry
  • Information Technology
  • Business Services
  • Startups

Telliscope is an AI-powered market intelligence platform built to help investors, researchers, and business professionals explore and analyze the African market. The platform aggregates and analyzes different types of data from multiple sources, including news, reports, statistics, events, and company profiles to uncover trends, identify opportunities, and provide actionable insights.

  • PostgreSQL
  • MongoDB
  • NestJS
  • GraphQL
  • ChromaDB
  • Sequelize
  • Mongoose
  • VertexAI
  • Gemini
4

Duration

Months

3

Team

Front-end Developer, Back-end Developer, Project Manager

1

Context

Telliscope allows users to discover new companies, follow industry trends, read important reports, and use AI search to find what they need quickly. At its core, the platform aims to bridge the information gap that exists in many emerging markets, with a focus on making African economies more accessible to global stakeholders.

To bring this ambitious vision to life, the Telliscope team partnered with us to build a data-rich, AI-enhanced web platform from the ground up. The product needed to deliver a seamless user experience, advanced backend architecture, third-party integrations and personalization features while remaining highly scalable and intuitive for its user base.

2

Challenge

Telliscope was a complex business platform requiring advanced backend architecture with a highly personalized user interface.

On the user-facing side, the platform needed to deliver a personalized dashboard that dynamically displays recommended companies, curated news, events, statistical data, and summarized reports- all filtered based on individual user preferences set during signup.

On the backend, we needed to build a powerful admin panel to support full content management for various data types. These included news (with RSS feeds and manual imports), event listings, report PDFs (processed with AI summarization), and statistical data sourced from external providers like the IMF. Another challenge was implementing an AI-powered search system that could understand user queries contextually and return meaningful, structured results from multiple data sources.

Moreover, the system required user management features such as email invitation, manual approval workflows, different permission levels, and analytics tracking through Google Tag Manager and internal logging. Given the heavy use of file processing, third-party APIs, vector databases, and AI models, the development required careful architectural planning and execution.

3

Solution

Our team began by designing a scalable and modular system architecture that supported both structured and unstructured data inputs. The tech stack included Nest.js for the backend, GraphQL as API layer, PostgreSQL and MongoDB for data storage, and ChromaDB for handling vector embeddings used in AI-based search. We also integrated Vertex AI’s Gemini models for advanced summarization and natural language processing tasks.

The frontend was designed with user personalization at its core. The dashboard was built as a grid-based layout where users could reorder or resize widgets to match their preferences. Personalized content was shown based on selected interests such as country, industry, or company, with bookmarking capabilities across all modules for easier access. Each dashboard component (e.g., news, reports, statistics) had its own dedicated detail page, and the AI-powered search allowed users to ask natural language questions and receive contextual, cited results.

On the admin side, we implemented complete CRUD functionality for all major content types. This included CSV/XLSX/PDF importers, RSS feed management, AI summarization tools for lengthy reports (including hover previews and full-view functionality), and metadata management including country/industry tagging and file storage in cloud services. User management was built with status toggling, invitation flows, and approval processes. We also incorporated dynamic chart visualizations and data exports for statistics.

Logging was added throughout the platform using MongoDB to track user behavior and preferences, which sets the foundation for future versions with even deeper personalization and analytics.

4

Development Process

The project was developed over four months, followed by a one-month testing and bug-fixing phase. The core team consisted of a front-end developer, a back-end developer, and a project manager who coordinated the workflow and maintained close collaboration with the Telliscope team.

Throughout the development, we held regular meetings to align on features, iterate on feedback, and ensure that both technical and business expectations were met. The tight feedback loop allowed us to rapidly prototype, test, and refine features like the dashboard layout, AI summarization accuracy, and the semantic search experience.

We paid special attention to performance optimization, given the platform’s reliance on multiple APIs, file uploads, and AI queries. Our infrastructure was deployed on Contabo, with Google Cloud services supporting file storage and AI functionality in production.

5

Result

Telliscope is now a powerful AI-driven business tool that helps companies find the information they need to make better decisions. It serves as a powerful tool for navigating complex markets with confidence, bringing together news, data, insights and intelligent search in one unified experience. The platform is currently live and fully operational, serving as a reliable source of market intelligence for the African investment landscape. It stands as a strong example of our capability to build complex, data-driven systems that combine AI and backend engineering into a great digital product.

Get an estimate

Describe your project by providing a written description, recording a voice message, or attaching a relevant file, your choice.