All projects
Side ProjectIndependent contributor

Juicebox

AI Recruiting Platform

Semantic search, candidate matching, and NLP features for an AI recruiting platform that helps recruiters find strong candidates faster.

PythonTypeScriptNode.jsReactNext.jsEmbeddingsVector SearchNLP
Visit juicebox.ai
juicebox.ai
Juicebox homepage: Win the talent war. Juicebox is the AI Recruiting Platform

Side project I contributed to in my own time, outside my regular employment. I didn't build Juicebox on my own, and my name may not appear on its public website. This page describes only the work I personally did.

01Project Overview

Juicebox is an AI-powered recruitment platform focused on candidate discovery and talent sourcing, helping recruiters move from a job description to a shortlist of strong candidates faster.

I worked on its search and matching features: semantic search that finds candidates by meaning, candidate matching and NLP, and automation that reduces manual sourcing work.

02Problem Being Solved

Keyword search misses strong candidates whose profiles describe the same skills in different words, and sourcing by hand takes recruiters hours. Search that understands meaning, plus automation of repetitive steps, gets recruiters to a good shortlist faster.

03My Role and Responsibilities

Side project: search, NLP, and full-stack engineering

I contributed to the platform's search and matching features, not the whole platform.

  • Created semantic search features that help recruiters find candidates by meaning rather than exact keywords.
  • Contributed to candidate matching and natural language processing (NLP) features for talent sourcing.
  • Automated repetitive recruitment tasks to reduce manual sourcing work for recruiting teams.
  • Built Python and TypeScript/Node.js services that generate embeddings and run vector search, with results surfaced to recruiters through React and Next.js interfaces.

04Technologies Used

Technologies I worked with in my contributions:

PythonTypeScriptNode.jsReactNext.jsEmbeddingsVector SearchNLP

05Architecture and Technical Implementation

Where my work fit

A simplified view of a candidate search. The highlighted layers are where my contributions sat.

  1. My contribution

    Recruiter search

    React and Next.js interfaces

  2. My contribution

    Search services

    Python and TypeScript/Node.js

  3. My contribution

    Embeddings & vector search

    Matching candidates by meaning

  4. Candidate shortlist

    Ranked matches for the recruiter

This is not a full map of Juicebox's architecture, which is proprietary.

06Key Features

Platform features

Features of the product as a whole, built by the wider team. My work covered semantic search, matching, and automation.

  • AI recruiting platform that learns how recruiters search.
  • Candidate discovery and talent sourcing from a job description.
  • AI agents that help source better talent, faster.

07Challenges and Solutions

Juicebox's codebase is proprietary, so implementation details aren't published here. I'm happy to walk through the problems I worked on in conversation.

08Screenshots and Demonstrations

juicebox.ai
Juicebox homepage: Win the talent war. Juicebox is the AI Recruiting Platform
Public homepage of juicebox.ai

09Live Application