LimiSearch · a managed service on LimiLake

Describe the assignment.Find the right people.

Describe the assignment in your own words. LimiSearch searches your consultants and candidates, and LinkedIn if you like. You get a shortlist with scores, reasoning and a link to each profile. We build and run the service, you use it.

How it works
  1. 1You describe the assignment in your own words
  2. 2We search your consultants and candidates, and LinkedIn if you like
  3. 3You get a shortlist with reasoning
LimiSearch · LimiLake

We need a frontend developer with three to five years of experience to cover parental leave in a design system team. Requirements: JavaScript/TypeScript, React, web components, preferably Lit, plus accessibility, CI/CD and component documentation. Desirable: WCAG/ARIA, focus behaviour, component APIs, Storybook, UI testing, design tokens and npm packaging. Close collaboration with designers and UX writers is essential. Who fits best, and which requirements lack supporting evidence?

We need a frontend developer with three to five years of experience to cover parental leave in a design system team. Requirements: JavaScript/TypeScript, React, web components, preferably Lit, plus accessibility, CI/CD and component documentation. Desirable: WCAG/ARIA, focus behaviour, component APIs, Storybook, UI testing, design tokens and npm packaging. Close collaboration with designers and UX writers is essential. Who fits best, and which requirements lack supporting evidence?

Profiles compared
TeamtailorCinodeBrainville

Five profiles to assess

  1. Emma LindbergExample score 92

    Senior frontend developer

    Lit and component APIs

    7 years · TypeScript, React, Lit, web components

    Project evidence
    Developed a Lit component library with properties, events and slots. Integrated components into React, wrote usage guides and set up component tests in CI/CD alongside the design team.
    What matches
    Strongest evidence for continuing Lit development and maintaining reusable component APIs. Documentation and collaboration with design are documented.
    What to follow up
    WCAG audits, manual screen-reader testing and Storybook are not documented. Collaboration with UX writers also needs confirmation.

    CV · Teamtailor · Example profile source

    Fictional CV preview for this demo. In use, a source link in the conversation opens the profile in your connected system, such as Teamtailor. No real profile is opened here.

  2. Johan NyströmExample score 88

    Senior frontend developer

    Accessibility and design system quality

    10 years · React, TypeScript, WCAG/ARIA, Storybook

    Project evidence
    Maintained a React design system with Storybook and design tokens. Documented component behaviour with designers and UX writers. Tested focus order, keyboard support and ARIA with NVDA and Playwright in CI/CD.
    What matches
    Strong evidence for accessibility, documentation and close design collaboration. A relevant alternative if the team can accommodate learning Lit.
    What to follow up
    Lit and web components are not documented. Verify this experience; the profile cannot be treated as a complete match.

    CV · Teamtailor · Example profile source

    Fictional CV preview for this demo. In use, a source link in the conversation opens the profile in your connected system, such as Teamtailor. No real profile is opened here.

  3. Sofia BergströmExample score 84

    Frontend consultant

    Web components in React

    6 years · TypeScript, React, web components

    Project evidence
    Integrated web components into a React application and documented events and properties. Worked on semantic HTML, keyboard navigation and automated tests in a cross-functional team.
    What matches
    Relevant experience using web components in React and helping other developers understand components.
    What to follow up
    Lit, Storybook and ownership of the component library are not documented. WCAG audits and collaboration with UX writers need verification.

    CV · Teamtailor · Example profile source

    Fictional CV preview for this demo. In use, a source link in the conversation opens the profile in your connected system, such as Teamtailor. No real profile is opened here.

  4. Erik LundqvistExample score 78

    Technical lead / full-stack developer

    Lit, packaging and versioning

    11 years · TypeScript, Lit, monorepo, npm

    Project evidence
    Built internal Lit components in a monorepo and published versioned npm packages through CI/CD. Documented APIs and supported several development teams.
    What matches
    Evidence for Lit, component APIs and library packaging. Relevant when the team needs to improve its technical workflow.
    What to follow up
    Previous assignments focused on broad full-stack and lead roles. Accessibility work, design collaboration and interest in a focused component role need follow-up.

    CV · Teamtailor · Example profile source

    Fictional CV preview for this demo. In use, a source link in the conversation opens the profile in your connected system, such as Teamtailor. No real profile is opened here.

  5. Linnea SjöbergExample score 73

    Frontend developer

    Documentation and design collaboration

    4 years · React, TypeScript, Storybook, design tokens

    Project evidence
    Contributed React components, Storybook examples and theming to a design system. Worked closely with a designer and UX writer on labels, error messages and usage guidelines.
    What matches
    Relevant to component documentation, design tokens and collaboration between design and development.
    What to follow up
    Web components, Lit, npm packaging and independent accessibility ownership are not documented.

    CV · Teamtailor · Example profile source

    Fictional CV preview for this demo. In use, a source link in the conversation opens the profile in your connected system, such as Teamtailor. No real profile is opened here.

An illustrative consultant search using Claude as one example interface through the LimiLake connector. An assignment is followed by a search of example sources and five fictional matches from an example network of 12,000 profiles. Open a row for supporting evidence and a fictional CV preview. Not a live search or an exact reproduction of the Claude interface. Timing is illustrative.

LimiSearch at a glance

  • Your consultants and candidates

    We search them for every assignment. LinkedIn is optional and shown separately.

  • Score and reasoning

    For each of your own consultants and candidates, with a link to the profile.

  • We run it

    Integrations, monitoring and updates are included.

  • Runs in Sweden

    On our own hardware and our own GPU servers.

A managed service on LimiLake

We bring profiles and CVs from your systems into LimiLake, our data platform. Parsing, search and AI assessment run there, and the shortlist reaches you in the channels you choose. We can build more services for you on the same platform later.

Architecture: three of your systems, for example Teamtailor, your CRM and Brainville, feed LimiLake. LinkedIn is searched for each assignment. LimiSearch parses, searches and assesses on Liminity's own hardware in Sweden and delivers the shortlist by email, in a dashboard or in AI assistants via MCP.

Teamtailor, CRM and Brainville are examples. We decide together with you which three systems to connect. Logos and trademarks belong to their respective owners.

From assignment to shortlist

Your profiles and CVs are ingested and parsed in advance. When you describe an assignment, LimiSearch searches, assesses the candidates and sends a shortlist.

  1. Continuously, in the background

    1Ingestion

    Profiles and CVs are pulled from your systems into LimiLake. New and changed profiles are picked up continuously, keeping the data current.

  2. Continuously, in the background

    2Parsing

    AI reads every CV, scanned ones included, and extracts skills, seniority, location and years of experience.

  3. For each assignment

    3Search

    We search on both meaning and keywords across your consultants and candidates, and LinkedIn if you like. Related experience counts, not just exact words.

  4. For each assignment

    4AI assessment

    Language models on our own GPU servers give your consultants and candidates a score and reasoning.

  5. For each assignment

    5Shortlist

    Your consultants and candidates, and profiles from LinkedIn, kept separate and linked to each profile.

Your profiles and CVs are ingested and parsed in advance. When you describe an assignment, LimiSearch searches, assesses the candidates and sends a shortlist.

Where you already work

The shortlist comes to where you already are. Pick one channel or several.

  • Email

    The shortlist lands in your inbox.

  • Dashboard

    All searches and shortlists in one place.

  • ClaudeChatGPTMicrosoft Copilot

    AI assistant via MCP

    Describe the assignment in Claude, ChatGPT or Microsoft Copilot and get the shortlist there.

MCP is an open standard for connecting AI assistants to other services. You connect with your own login. It also works with your own AI agents.

If the shortlist is sent to an external service, such as email or an AI assistant, that service's terms apply to the content. We record the channels you use in the data processing agreement.

ClaudeChatGPTMicrosoft CopilotYour AI assistant
LimiSearch
Via MCP, with your own login

Operations and security

  • Runs in Sweden

    Ingestion, OCR, search and AI assessment run on our own hardware in Liminity's Swedish data centre.

  • No external AI for CVs

    CV text is never sent to an external AI service.

  • DPA included

    A data processing agreement is included in the service.

Each customer's data is kept separate from other customers', and stored credentials for your systems are encrypted with customer-specific keys.

What's included

Onboarding and integrations are included. Contact us and we'll go through your systems and assignments and send you a quote.

  • Three integrations

    Tailored to your needs, for example Teamtailor, your CRM and Brainville. More can be added later.

  • LinkedIn search

    Every search can also cover LinkedIn. Those profiles are shown separately.

  • Operations on LimiLake

    We run, monitor and update the service on our own hardware in Sweden.

  • Unlimited searches

    Search as often as you like. The shortlist arrives by email, in the dashboard or in your AI assistant.

How to get started

  1. 1

    Review

    We go through your systems and choose together which three to connect.

  2. 2

    Integrations

    We build the integrations and ingest profiles and CVs.

  3. 3

    Delivery

    We set up the channels you chose: email, dashboard or AI assistant.

  4. 4

    Search

    You describe the assignment and get the shortlist.

Want to get started or learn more?

Get in touch, and we'll start with a short review of your systems and assignments.

LimiSearch FAQ