Book a strategy call
Back to case studies
Legal Services, AI
September 7, 2026

CaseCraft.AI: Building the Future of Small Claims Resolution in the UK

How a legal tech startup built a complete engineering function consisting of a CPTO, Backend and AI Engineer and two Full-Stack Engineers.

The Future of Small Claims Resolution in the UK
The Future of Small Claims Resolution in the UK

Challenge

CaseCraft.AI is a London legal tech company building small claims resolution for the UK. Small claims is a high volume, low value process, and the people using it are both individuals and businesses. That sets two hard limits at once: the product has to be accurate enough that someone will trust it with a real dispute, and cheap enough to run that the unit economics survive. Automating a legal workflow to that standard is a complex engineering undertaking, and the company needed it with no engineering organisation of its own.

There was no technical leadership in the business. Nobody owned the architecture or the build-versus-buy decisions that would determine what the product cost to run, and no hands-on engineering team sat behind that gap. Each role was a rare combination rather than a single skill. The CPTO role needed distributed systems design, data engineering and LLM applications from legal domain shipped into production: Python, PyTorch, Spark, Databricks, AWS and LangChain. The Backend and AI Engineer role needed high-load microservices on Java, Spring Boot and Kafka alongside production OpenAI and vector search experience. The two Full-stack Engineer roles needed commercial React, Vue, Node.js and TypeScript, plus the judgement to own the frontend and the services behind it.

We took the leadership request first, on the principle that this hire should be founding, and that engineers working under a technical leader ought to be chosen with that leader in the room. We sourced and screened against full requirements and experience in the industry rather than the job title, and the founders interviewed and made the decision themselves. The CPTO was hired 74 hours after the request. The three engineering requests were then created against the architecture that CPTO had settled on, rather than guessed at before anyone knew the stack, and all three engineers were placed behind it. The engagement has been running for over 18 months with the whole engineering team still intact.

x4
Talents Secured
18+
Months Duration
74h
Time-to-hire

Outcome

We placed all four roles. The CPTO came first, within 74 hours of the brief, and the backend and full-stack hires followed once there was someone in post to define what the team around them needed to look like. The founders interviewed and decided on every one of them.

  • Technical leadership in place 74 hours on the role that usually takes a startup the longest.
  • A complete engineering function at a company of under ten people: architecture and data, AI infrastructure, and full product delivery.
  • All roles are filled with no internal recruiter overheads or separate agency per hire.
  • Over 18 months, all engineers are still on the engagement rolling out new features and workflows.
Yotewo helped us a lot, brought loads of candidates and we were very happy to find our current CTO. The model they operating is very efficient for startups.
Alexander Mints
Alexander Mints
Co-Founder, CaseCraft.AI

A Strategic, Multi-Phase Hiring Approach

Phase: Request

Four requests defined by stack, seniority, and industry experience: distributed systems and production LLM work for the CPTO, Kafka and vector search for the Backend and AI Engineer, commercial React and TypeScript for the two Full-Stack Engineers.

Phase: Matching

Profiles of 12 engineers were shortlisted across the four requests and 6 went to interview, each one vetted before the founders saw them.

Phase: Interviews

Interviews were booked with no limit on how many, every decision stayed with the founders, and the first engagement was confirmed 74 hours after the request was opened.

Phase: Engagement

All four engagements opened with a monthly full-time workload, daily tracker, and monthly reports with breakdown of hour commitments.

Phase: Support

As CaseCraft.AI grew, workloads moved up and down, extra hours were requested, and across four engineers over eighteen months every change was recorded and approved in the account.