Full-stack · Vocab.AI · Aug 2023 — May 2024 · React · Llama 2 · GPT-3.5 Turbo
Conversation analytics
Customer-service calls turned into something a team can act on: who spoke, how they sounded, what it was about, and how it went.
What I did
As a full-stack developer intern I built speaker diarization and voice emotion recognition models and the React app people used to run them, fine-tuned Llama 2 and GPT-3.5 Turbo for sentiment, topic modelling and conversation scoring, and wired an automated QA system and REST APIs into customer-service workflows.
Who I made it with
The Vocab.AI team.
What it’s for
Vocab.AI builds conversation analytics for customer-service teams.
Why it was hardThe models were only half of it. The analytics had to reach the people running support through an app and APIs they would actually use.
Receipts
- Previous portfolio, as it stood before this redesign (self-reported)Self-reported: IIIT Dharwad (Data Science & AI), research at IIT Bombay Educational Technology, CRF approach for SSMR.
Next, a little further back in timeEmotion detector for online classes →