My role: System Architect & Builder

I built a job search workflow for someone in the Netherlands who needed to find suitable vacancies quickly during her zoekjaar period and focus only on companies that could realistically hire her through the kennismigrant route.

Instead of manually searching LinkedIn, checking every employer against the IND recognized sponsor register, and copying vacancies into a tracking sheet one by one, she can now trigger the process with a simple Telegram message and review a ready-made list of filtered roles.

Video walkthrough: Watch on YouTube

The goal

The process needed to reduce the most repetitive parts of the search:

• finding newly posted vacancies on LinkedIn
• checking whether the employer appears in the IND recognized sponsor register
• avoiding duplicate vacancies across repeated searches
• writing structured vacancy data into a spreadsheet for later application tracking

This matters because people in this situation often need to send a high volume of applications, and the sponsor check alone adds a large amount of manual overhead.

What I built

I designed a Telegram-triggered n8n workflow that starts from search parameters stored in Google Sheets.

The sheet acts as the control layer for:

• search keywords
• the number of vacancies to collect
• the time window, such as the last 24 hours or even the last hour

Once triggered, the workflow:

1. reads the search configuration from Google Sheets
2. loads previously collected vacancy IDs to avoid duplicates
3. runs a LinkedIn vacancy search through Apify
4. normalizes and filters the results
5. compares employer names against a local copy of the IND recognized sponsor register
6. writes new vacancies into the spreadsheet for review and application tracking

System architecture

The workflow combines three practical layers.

1. Vacancy discovery

The search starts from a role keyword such as Operations and requests vacancies posted within a recent time window.

This makes the system useful for fast-response job search, where timing matters and older vacancies are less valuable.

2. Duplicate control

Each vacancy has its own ID.

Before processing new results, the workflow loads the IDs that have already been stored in the spreadsheet and excludes them immediately. That keeps the output clean across repeated runs.

3. Sponsor verification

I copied the full IND recognized sponsor register, around 12,000 companies, into a spreadsheet and used that as the comparison layer.

For each vacancy, the workflow tries to match the employer name against that register and records the result:

• 1 if the employer appears to be a recognized sponsor
• 0 if no useful match is found

In most cases, the matching works well and removes a large amount of repetitive manual checking.

There are still edge cases where string matching can produce false positives, so the final shortlist is reviewed visually before applications are sent. That tradeoff is still far more efficient than checking every company manually from scratch.

Results

The system works as a practical filtering and collection layer for a difficult job search process.

Operational outcome

• a single request can return up to 200 vacancies
• only new vacancies move forward into the spreadsheet
• obvious non-matches are filtered out early
• sponsor verification is handled automatically for most cases

Workflow outcome

Instead of collecting vacancies manually and then checking each company one by one on the IND website, the user now starts with a structured shortlist that is already enriched with sponsor-match signals.

That means more energy can go into reviewing roles, tailoring applications, and sending resumes instead of doing repetitive research and copy-paste work.

What can be added next

This workflow already stores the full vacancy descriptions, which makes the next automation steps straightforward.

Two useful extensions are:

• match each vacancy against the candidate’s resume and score fit
• generate draft motivation letters based on previous applications, the resume, and the vacancy text

That would turn the system from a vacancy collection workflow into a semi-automated application preparation pipeline.

My role

I designed the workflow logic, structured the spreadsheet model, built the search and deduplication flow, prepared the sponsor verification layer, and packaged the process behind a simple Telegram trigger so it was easy to use in day-to-day job search.

Tech stack

• n8n
• Apify
• Google Sheets
• Telegram
• OpenAI API
• Claude
• LinkedIn job search
• IND recognized sponsor register