News from CodersFirst R&D department
We think that all companies, from early-stage startups, through scaleups, to mature tech companies should have the possibility to focus on their core mission to achieve KPIs without any distractions. To make it possible, since the day CodersFirst was established, we have been working on the perfect recruitment process for the software engineering industry to make your life easier and enable you to skip repetitive sourcing, screening and tech verification of each potential candidate.
In our R&D department, we have designed a blinded and skills-focused recruitment process to screen all engineering abilities and validate candidates’ experience in a natural coding environment. This allows us to identify candidates’ strengths and predict where the programmer can add the most value and spread wings. Having this knowledge we are able to match software engineers with companies looking for a specific skill set and abilities.
As a bonus, this formula is entirely stress-free for both sides (programmers and companies), bringing our scoring to the highest level of accuracy (according to research).
We think that our solution is a huge time and money saver, as we introduce you to better software engineers with the biggest chance for employment at your company. And science agrees with us.
Our process consists of two parts:
- A screening test with a coding task
- A structured live-coding interview
Screening test + coding task
This test assesses candidates’ engineering strength by screening programming skills in a number of different domains, such as databases, architecture, program optimization, web development, etc. You can find here mostly questions with lines of code. The adaptive approach means that each candidate has different questions. Questions are updating in real-time based on a candidate’s previous answer and selecting by taking into account a candidate’s current skill level. By using this system we are able to screen a whole spectrum of knowledge in computer science and software engineering. The mechanism is similar to the mechanism used in GMAT and SAT tests.
Immediately after the test, software engineers are asked to complete one simple coding task.
In this part, we ask candidates to make use of our dedicated editor and solve the programming task in a programming language of their choice. The tasks are simple for a reason, as according to research and our tests this is the only way to check the logic behind given solutions. We also check actual programming skills and assess code quality. We take results from this part of our process into account while deciding if candidates should be invited to the last part, which is a live-coding session. For this part, we pick only the most promising candidates who have the skills needed at tech companies.
Structured live-coding interview
Here, our trained technical recruiter conducts a live-coding session in a language of candidates’ choice using the rules of a structured interview. This part is highly structured and exactly the same for all candidates (to see why -take a look at one of our previous articles about structured interviews). It is always performed in candidates’ natural working environment (on their own computer) so that they can feel most comfortable. Having collected data from all parts of our process, we can match candidates with companies they fit best with.
How do we do that?
What do we do with that insane amount of data?
Data analysis
We are continuously collecting and combining more and more data and conducting more and more recruitment campaigns. This allows us to fine-tune our entire process. We already found some interesting results. While data collection is an endless process, we decided to share some of our findings and updates with you every time we identify a useful discovery.
Recently we found out that our first step nicely predicts whether candidates will perform well in the next parts of our recruitment process. This means that our adaptive test is identifying skills well.
Secondly, 52% of Final Interviews through CodersFirst result in an offer, while the industry standard is 20%. We are proud because we want to be able to predict the likelihood of a programmer getting an offer from a particular company and statistic says we’re on track.
Additionally, we are confirming that our simple programming task and the structured live-coding interview correlate well with job performance. In theory, they should, since work sample tests and structured interviews are great predictors of job performance. However, for all those analyses to be feasible we still need more long-term data. Thanks to science we know what to expect, but we always need to check everything on our own to be absolutely certain that what we are doing makes sense.
With all these exciting results we can really give valid recommendations to our clients. We also want to use machine learning algorithms to improve our data analysis and candidates-employers matching. So as you can see we are constantly improving, changing and expanding our recruitment process, which is already pretty good! We will keep you posted on what CodersFirst is currently doing on a regular basis, so stay tuned for further updates and feel free to contact us at any time!
Special thanks to Aleksandra, our co-founder, and COO, for her input to this article.