AI pairing engine — made in Switzerland

The right pairs,
right now.

A light, safe and configurable matching engine. You set the criteria and their weights — PAIRS forms the pairs that maximise overall relevance.

Explore pricing How it works

+Proven on 285 pairs at EPFL +LPD / GDPR compliant +Hosted in Switzerland

01 — About

Pairing by hand takes days.
PAIRS does it in hours.

The problem

  • Matching mentors and mentees by hand takes days — and scales poorly.
  • Criteria are many and hard to weigh: background, sector, goals, languages, availability.
  • Existing tools are pricy, not privacy-safe, or reserved for large organisations.
  • The result: sub-optimal matches and heavy follow-up.

Our response

PAIRS turns matching from an Excel bottleneck into an automated, traceable step.

  • Weighted criteria, adaptable to each program — dimension by dimension.
  • An optimisation engine that maximises the overall matching score, exact matches and preferences included.
  • Validation and manual changes directly on the platform, with ready-to-use exports.
  • Freedom in choosing the most adequate language model.

Case study — EPFL Alumni Mentoring Program, 9th edition

Where PAIRS was born and proven. Record participation.

0pairs formed
0participants — 287 mentees, 182 mentors
0%would recommend the program
0%find it useful or very useful

2025 participant survey · A second case is already on its way: preference-based matching, where each person submits three choices.

The logic goes beyond mentoring.

Wherever the right pair changes the outcome.

EPFL programs

Forum, MTE, VPI, Philanthropy — the first showcase.

Universities & alumni

Alumni offices and academic mentoring programmes.

Companies & HR

Onboarding, buddy programmes, internal mentoring.

Incubators & accelerators

Matching startups, mentors and experts.

Clubs & associations

Matching members across small structures.

Events & networking

Speed-meetings and themed matching.

Built by a pair, naturally.

Born at EPFL, engineered in Lausanne.

Yann Vallotton

Tech & development

Designs and builds the platform and the matching engine. System architecture & digital marketing.

Jordane Widmer

Product, operations & program

Leads the product, the programs and the relationship with users. 4+ years in communications, MSc in Economics.

Sovereignty first. Matching data is personal data. Processing and hosting in Switzerland — LPD & GDPR compliant — via EPFL infrastructure or trusted Swiss partners, with full freedom in choosing the language model.

02 — Pricing

Dimension by dimension.

A simple model: two plans, charged per user. The difference is how deep your matching goes.

Basic

For small pairing needs

Per userbilled per campaign

  • Simple forms, few dimensions
  • Weighted criteria & scoring engine
  • Validation and manual changes on the platform
  • Excel / CSV import and ready-to-use export
  • Swiss hosting, LPD / GDPR compliant
Scope your campaign
Most capable

Horizon

For important pairing needs with complex data input

Per userbilled per campaign

  • Complex forms, high-dimensional profiles
  • Preference-based matching — each person ranks choices
  • Advanced weighting, traceability & exports
  • Alumni-platform integrations, API in time
  • Priority support & campaign reiteration savings
Talk to the team

Both plans run on sovereign Swiss infrastructure — and you keep the freedom to choose the most adequate LLM.

03 — Contact

Let's form the right pairs.

Tell us about your program — cohort size, criteria, timeline. We answer within two working days.