Skip to content
Deependra Vishwakarma
Start a project

2024 · B2B SaaS platform · Exhibit Magazine India Pvt. Ltd. · Team lead and architect

Exhibit Social

A B2B influencer-marketing platform built from zero: 10,000+ influencers indexed and 30+ brand campaigns run on it.

I led the team that built a B2B influencer-marketing platform end to end: influencer discovery, automated data collection, analytics from the Meta and YouTube APIs, campaign management and a collaboration workspace for brands and creators.

10,000+

influencers indexed

How this is measured

Influencer profiles collected and kept up to date by the platform's pipelines.

30+

large brand campaigns

How this is measured

Large brand campaigns run on the platform.

Screenshot of Exhibit Social
The live product at https://app.exhibit.social/.

The constraint

Brands wanted a reliable, data-driven way to find and vet influencers for campaigns, instead of spreadsheets and guesswork.

Discovery, real-time analytics, automated data collection and a collaboration workflow, all built from nothing, on top of third-party APIs with strict rate limits.

The architecture

  • Automated collection pipelines in Python gather and store influencer metrics on a schedule.
  • Rate-limit-aware API clients for Meta and YouTube, with request queues and exponential backoff.
  • A Laravel and Vue.js platform for discovery, campaigns, analytics dashboards and brand–creator collaboration.
  • MySQL and Redis on Google Cloud for storage and caching.
Architecture of Exhibit SocialScheduled Python pipelines collect influencer metrics from the Meta and YouTube APIs through rate-limit-aware clients and store them in MySQL; the Laravel and Vue.js platform serves brands from MySQL with Redis caching on Google Cloud.rate-limited, with backoffrate-limited, with backoffmetricsqueriescachediscover, run campaignsMeta Graph APIYouTube Data APICollection pipelinesPython, scheduledMySQLGoogle CloudPlatformLaravel and Vue.jsRediscacheBrands
the hot path asynchronous
The diagram as text

Scheduled Python pipelines collect influencer metrics from the Meta and YouTube APIs through rate-limit-aware clients and store them in MySQL; the Laravel and Vue.js platform serves brands from MySQL with Redis caching on Google Cloud.

  • Collection pipelines → Meta Graph API: rate-limited, with backoff (asynchronous)
  • Collection pipelines → YouTube Data API: rate-limited, with backoff (asynchronous)
  • Collection pipelines → MySQL: metrics
  • Platform → MySQL: queries
  • Platform → Redis: cache
  • Brands → Platform: discover, run campaigns

Trade-offs

  • Scheduled collection over on-demand lookups: data is minutes old instead of live, and the platform never trips API limits.
  • One team owning every layer, from data pipelines to UI, meant less coordination and faster shipping.

Results

  • 10,000+ influencers indexed and kept up to date automatically.
  • 30+ large brand campaigns run on the platform.
  • Influencer data is collected automatically instead of by hand.

What I'd do again

  • Automated data collection is the backbone of a B2B data product. Manual entry doesn't scale.
  • Design for third-party rate limits from the first day. Queues and backoff prevented account suspensions.
  • End-to-end ownership shortens the path from idea to production.

Stack

  • Laravel
  • Vue.js
  • Python
  • MySQL
  • Redis
  • Meta Graph API
  • YouTube Data API
  • Google Cloud

Related services: MVP Sprint, Web platforms and SaaS, Architecture review and due diligence, Fractional CTO.

Need something like this?

A short brief is enough to start. I’ll reply with questions, a suggested first step and when I could begin.

Start a projectSee the services