# Case Studies — 20Hypes

> Delivered work with the challenge, what was built, the technical contribution, and the outcome.

Canonical page: https://20hypes.com/case-studies

## Document Intelligence Platform

An AI-driven document processing platform that extracts and classifies data from scanned and digital documents at more than a million file transactions a month.

**Type**: Engineering contribution

**Focus**: AI document processing, full-stack engineering & cloud infrastructure

**Result**: 1M+ file transactions / month

**Role**: Full-stack engineering, Cloud infrastructure, System architecture

**Technologies**: AWS, Node.js, TypeScript, OCR & AI extraction, Queues & events, S3, SharePoint API

**Document processing pipeline**: Documents -> API -> Queue -> AI Extraction -> Classification -> Delivery -> SharePoint + Webhooks

### Challenge

Documents arrived as both scans and digital files, often as large PDFs containing many separate documents bundled together. Reading them, pulling out the required fields, working out what each document actually was, and filing it in the right destination was manual work that no team could keep up with at volume.

### What we built

An AI-driven processing platform that goes well beyond OCR: ingestion APIs, text and data extraction from scanned and digital documents, extraction of the specific fields each document type requires, classification and splitting of large PDFs into their constituent documents, and outbound delivery into systems such as SharePoint and customer webhooks.

### Technical contribution

Work spanned application and infrastructure: building the processing and extraction pipeline and its product features, then engineering the capacity, scalability, and reliability needed to keep the platform predictable as volume grew.

### Outcome

The platform sustains more than one million file transactions per month, with documents classified and routed to their destination systems automatically instead of by hand.

## AI Code Generation Engine

An AI-driven code engine that generates frontend and backend applications from structured requirements.

**Type**: Engineering contribution

**Focus**: AI engineering & platform engineering

**Result**: 7,500+ APIs generated in ~2–3 minutes

**Role**: AI engineering, Platform engineering, API generation

**Technologies**: TypeScript, Node.js, LLM APIs, Schema-driven APIs, Code generation, PostgreSQL

**Generation pipeline**: Requirements -> AI Code Engine -> Backend + Frontend -> APIs

### Challenge

Every new project repeated the same foundations — data models, endpoints, validation, and screens — and writing that layer by hand did not scale with the number of projects.

### What we built

A code engine that turns structured requirements and schemas into working applications: generated backend APIs with validation, matching frontend application code, and the pipeline that runs generation end to end.

### Technical contribution

Engineering across the generation pipeline: the schema-driven API generation model, the AI-driven generation of application code, and the checks that keep generated output consistent and usable.

### Outcome

More than 7,500 APIs were generated in approximately two to three minutes, turning a repetitive build phase into a generated one.

## Hireon

A recruitment platform that processes job descriptions, analyses CVs, and scores candidates against a role.

**Type**: Product

**Focus**: Recruitment technology & AI

**Role**: Product architecture, AI engineering, SaaS development

**Technologies**: Next.js, TypeScript, Node.js, PostgreSQL, LLM APIs, AWS

**Recruitment workflow**: Job -> AI Analysis -> Candidates -> Matching -> Ranking

### Challenge

Hiring teams work across fragmented tools, and reviewing each candidate against a role is slow, repetitive, and inconsistent between reviewers.

### What we built

A SaaS recruitment platform covering job and application workflows, job description processing, CV and candidate analysis, and AI-assisted matching that scores and ranks candidates against a role.

### Technical contribution

Built as an independent 20Hypes product: product and data architecture, the AI analysis and scoring pipeline, and a modern SaaS application around it.

### Outcome

Hiring workflows and candidate review run in one product, with structured, repeatable scoring in place of manual CV screening.

## Vida

A Flutter-based, ad-free video player built for a clean, uninterrupted viewing experience.

**Type**: Product

**Focus**: Mobile application development

**Result**: 0 ads or interruptions

**Role**: Flutter development, Mobile architecture, Cross-platform UI

**Technologies**: Flutter, Dart, Video playback, Cross-platform

**Playback flow**: Video -> Flutter App -> Ad-free Playback

### Challenge

Most mobile video players interrupt playback with advertising and clutter the interface with everything except the video.

### What we built

A cross-platform video player built in Flutter, with a deliberately minimal player interface, smooth playback controls, and no advertising anywhere in the experience.

### Technical contribution

Built as a 20Hypes product: Flutter application architecture, playback and player integration, and a single codebase that behaves consistently across platforms.

### Outcome

One Flutter codebase delivers the same clean, ad-free playback experience across mobile platforms.