Zij - Details
relevant skills
- 1) exploratory data analysis with sql
- 2) predict from data with python
- 3) software design patterns
- 4) statistical foundations
- 5) version control with git
- 6) non-linear thinking
- 7) local language proficiency
1) exploratory data analysis with sql
- You are a data scientist in a health insurance company. In your firm, claims experts need to manually review claims.
- A business domain expert wants to find out whether you could set up an impact project together. Here are two of her impact ideas:
- Some claims handlers might have expert knowledge that they could teach to the rest.
- Some claims might be forwarded more often than necessary before the claims review is completed.
- Using the following synthetic sqlite database:
- Spend 30 minutes to explore the data. You can save queries and corresponding results in "sql-explore.sql".
- Afterwards, present relevant characteristics to the domain expert and expand your analysis together.
- Finally, develop a project proposal together.
2) predict from data with python
- You are a data scientist in a health insurance company. In your firm, claims experts need to manually review claims if an automated, rule-based system identifies a reason for review (manual_review = 1). Every claim has some defining parameters such as claims amount or treatment date (features.csv), as well as information on whether a claims expert was able to reduce the claims amount (reviews.csv).
- Your job is to predict the economic impact of a manual review on an incoming invoice. In preparing your results, please
- explore the features and their relationship to the claims result,
- preprocess your data to improve model performance,
- train a baseline model to predict the economic impact of a manual review.
3) software design patterns
- use pre commit hooks (esp. for linters)
- write unit tests & integration tests
4) statistical foundations
- You are a data scientist at an ecommerce store for second hand sneakers. Up until today, experts at the company have been setting prices manually. Going forward, it is your task to set up a pricing algorithm to set prices automatically.
- The company management asks you to come up with a) a list of things you would like to know about each item and b) a rough outline how you will use this data to predict prices.
5) git challenge: https://github.com/juanfresia/git-challenge
6) non-linear thinking
- why is your current job relevant?
- what do you do on a typical day?
- which leverage points do you have?
- how do you evaluate whether you're making progress?
- which metrics are relevant in your job?
7) local language proficiency
- get at least B1 in your target language
USPs for companies:
- Traditional recruiting agencies connect candidates with companies - and nothing else.
- In contrast, we train candidates and coach them - even after placement. They remain lifelong members of our community for continuous learning.
Role model: a16z
- a16z saw that traditional top-tier VC companies provided mostly capital and personnel (via board seats) as USPs.
- In contrast, they were able to disrupt the VC market by - among other things -
- a) betting on founders (while others exchanged C level personnel in times of crisis),
- b) establishing standardised networking opportunities (e.g. B2B or industry-specific networks)
Potential business model: build a recruiting network for non-linear skills
- Placement fees: Receiving companies pay 30% of the candidate's annual salary if the candidate isn't terminated within 6 months.
- Consulting fees: Candidates are initially placed as consultants at interested companies. If a company takes in the candidate, incurred fees are deducted from the placement fee.
Long-term strategy:
- Build a healthcare service provider that helps fix Europe's structural healthcare problems.
- Build a global vocational education school to replace existing shallow, broad higher education.
Long-term macro trends:
- More developed countries face an increasing shortage of talent, which increases employers' hiring & retention costs.
- Less developed countries have a younger population that is looking for work opportunities abroad.
Here is LLM-generated feedback during a recertification conversation:
