The Intern Group
- HR Tech
- Recruiting
- Talent Acquisition
- Early Talent
- Talent Management
- Employability
- Talent Matching
- Conversational Search
- UX/UI Redesign
The client
The Intern Group connects students and graduates with internships around the world. Their network includes more than 3,000 partner companies, and over 2,000 participants join their programs each year.
The challenge
With thousands of openings across many companies and career fields, finding the best opportunity for each candidate took a lot of time. The existing matching algorithm relied on broad filters and returned results without a meaningful ranking, so the best options were often buried deep in the list.
To make a good match, the team had to go through results one by one and review profiles and openings by hand. This slowed placements and made the process hard to scale as the network grew. The team needed a smarter way to bring the right opportunities to the top.

The solution
We designed and built an AI recommendation engine. For each candidate, it delivers a ranked shortlist of the best-matching opportunities, and every result comes with a clear match score.
The model was trained on the company’s own history of successful matches, so it reflects how their experts actually make decisions. It understands what a candidate wants to achieve, not just which keywords appear in their profile. When the results need adjusting, the team can refine them in plain language and the system returns new recommendations.
We also redesigned the matching experience from the ground up. The new interface puts the best opportunities front and center, shows each match score at a glance, and lets the team review, compare and act on recommendations.
The value
Faster matching
A ranked shortlist is ready in seconds, which replaces hours of manual searching.
A smoother workflow
A redesigned interface turns a long, manual process into a few clear steps.
Recommendations people trust
A score on every result makes each suggestion easy to understand and act on.
Better-fit matches
The system weighs aspirations and context as well as keywords.
Full catalog coverage
Every relevant opening is considered, not just the first results.
Built-in business rules
Hard requirements such as location, language and availability are applied automatically.
A system that keeps improving
Every confirmed match makes the model smarter over time.
Higher conversion across the entire process
AI-suggested matches convert better at all four stages of the funnel, from initial interest to role confirmation. The improvement ranges from 10.6 to 13.3 percentage points compared with the traditional process.
Better matches from the first touchpoint
The biggest advantage comes at the very first step: more suggestions generate interest in an interview, and that carries through the rest of the process to placement.
How it works
- 01
Prepare the data
Incomplete or invalid listings are cleaned out so the model only learns from reliable information.
- 02
Learn from past decisions
A custom AI model is trained on historical matches and learns which factors matter most.
- 03
Understand profiles and opportunities
Candidates and openings are translated into a shared semantic space, so similarity is measured by meaning rather than by exact keywords.
- 04
Rank and filter
The engine finds the closest matches, scores them, and applies the client’s business rules.
- 05
Review in a redesigned interface
Recommendations are shown in a clear, ranked view with match scores, so the team can decide quickly.
- 06
Refine with natural language
Users describe what they’re looking for in their own words, and an AI agent adjusts the search.
- 07
Improve continuously
Real outcomes feed regular retraining, with safe version control and rollback.

Where Gleni can help
Next case
Scottish distilleries, with Ardnamurchan Distillery and Isle of Harris Distillery among its first users in production. These operations combine traditional craftsmanship with modern regulatory demands: hundreds of casks maturing in the warehouse, dozens of distillation batches per month, and the obligation to report every litre of alcohol to HMRC, the UK tax authority.
- Proprietary SaaS
- ERP
- Traceability
- Scotch Whisky
- Multi-tenant
- Compliance
SaaS ERPExplore work