The brief
Extend Enterpret's unified AI feedback platform by building production-ready integrations for multiple new feedback sources — CRM, review platforms, social media, video conferencing, and file-based ingestion — following the existing Golang microservice architecture pattern, with each integration delivering OAuth flows, data fetchers, transformers, and normalization into Enterpret's singularity feedback record model.
What we built
A comprehensive multi-channel feedback ingestion platform expansion for Enterpret, delivering 10+ production-grade integrations in Golang: HubSpot (feedback + user/account sync with custom objects), Medallia, Trustpilot (multi-language with historical backfill), Zoom (standard API + real-time RTMS v2 WebSocket with PCM audio, transcription, and S3 pipeline), SFTP (TrustVoice and custom providers with streaming JSON/JSONL/ZIP parsing, byte-offset resume, and TOFU security), Slack (multi-tenant webhook fan-out + app store approval), YouTube, Twitter DM, Meta Threads (with reply-chain flattening), and Gong. Each integration follows a consistent fetcher → transformer → singularity pipeline pattern with OAuth2, token management, rate limiting, watermark-based incremental sync, and GraphQL/Protobuf contracts. The team also maintained continuous multi-environment deployment operations across 5+ microservices.



