Well all was good but since we were into hiring from a long time, We gathered a fair list of candidates interested (through email, linkedin hiring posts, etc), checked in through some sanity to see which candidates to reach out first but even then, We realized some major bottlenecks:

1. Conducting basic screening for all the candidates were very labour intensive / time consuming.
2. Often we could find the good vs bad applicant at a glance after talking to them for 5 mins, while similiarly a resume-based shortlisting made us miss out some really great talent.
3. Time taken to conduct all screening, making sure we do not accidentally filter out a good talent, made us waste dozens of hours with those who could be screened early on!
4. In case of multiple Interviewers, there was always a bias and explaining why they felt a candidate “felt better” than the other one was difficult, we needed structured/summarized report.
5. Related to no.4, writing a summary of each interview was time consuming.. Already piling up on limited time the team had for technical velocity and moving fast.

If you notice, We faced that typical problem any SMB face, Hiring based on “gut feeling”, which we felt (after a good amount of hits and trials to be wrong)

Well, as a technical guy and someone who believes in LLMs/AI revolution, I took it upon myself to find maybe a better/optimized solution, which can perhaps solve some (if not all) of our pain points.. and that is when we found Althire AI

Introducting Althire AI — End to End Recruitment Solution for Everyone

Fast forward some research and exploring their page and they do, We found out Althire as a recent emerging startup in HRTech, focusing on similar problems I mentioned earlier (Never in my life, I felt so targeted)

At time of writing this (26 March) they had a couple of Interesting offerings, but what Intrigued us the most was their product TalentBoost (https://althire.ai/products/TalentBoost) which, as mentioned is end to end interviewing platform with basic ATS, Since we were not really sure if they can help our problem, We reached out to Althire at contact@althire.ai, surprisingly a reply came in fast and a demo was scheduled!

Fast forward a meeting with some of the greatest founders we have met (they were very humble, friendly), They agreed to let us do test runs and pay only if we see “it’s actually worth it”! Infact since it is launched quite recently, they were offering free access (and it is all self-serve, without any complicated meetings/request for access, etc!) for some time (I hope you can avail the same, act fast!)

Now you must be thinking, why am I advertising so much and praising them in such good lights? Well, because they deserve it!

Don’t believe so? I’ll be showing you the exact steps we took and how we used their platform to do all the screening part (keep on reading!)

Using Talentboost — Step by Step

First we went to their dashboard, signed up using company account (important), and were presented by this screen:

Here I have marked some important regions, clicking on the button in green will have us create a new job (basically design an AI interview for screening candidates)

Filling in the field and clicking on Create Manually, It asked for a job description (which we quickly filled in)

After adding on a job, we were presented with this screen:

Now you might ask, what is “Craft Interview” ? Well! TalentBoost gives an option for us to manually cross-verify the Interview script and modify/change how to evaluate a good candidate, or the questions that need to be asked!

Since the “Crafting of Interview” itself is a big topic, I’d like to break it down (as to how we approached it when using the platform!)

In the above, the most impactful to us was:

Where each of the highlighed ones are the different dimensions/aspects a candidate is evaluated on (ie in this example, a candidate good in social media strategy can get a score of 10/10, but he might be bad in content creation and score 4/10) — Evaluation is not one dimensional (ie single score), but rather multiple dimensional (A candidate be good at one vertical but bad at another!)

and finally, the most important one:

Where in brown we have a main question (imagine like the theme of what you are going to ask, If taking an in-person interview) and a set of possible followups that the AI interviewer can ask!

Modifying the above interview and clicking on “Confirm”, We got the option to “Add Candidate”.. (Well, let’s do it!)

Finally, I (as a candidate) will get this amazing email:

Once the candidate clicks on the link, they will be redirected to Althire’s page where their agent will conduct end-to-end interview without any intervention! Once finished, they even generate a very structured report :)

Similiarly, talentboost basic ATS provide you option to filter/shortlist candidates like:

Here is a sample report:

Our success with the platform

We have rigorously put this software at trial, though there are some minor issues/bugs that pops in here and there, our overall experience has been nothing but amazing

Cost Savings? Where are the numbers??

Here is a rough estimate of how useful this tool (and any good LLM/AI based tool could be in emerging HRTech) to us was:

In Average before Talentboost:

• Each candidate used to take around half an hour per interview
• 10–20 Interviews were the maximum we could afford to conduct (in a day) (Per person)
• We used to interview maximum 20–80 people before deciding on a hire
• As a result, quality of hire was regressed, the manual process of conducting interview became groggy and the hired talents were “so-so” than being a top talent
• Used to miss out a bigger pool, hundreds of more candidates interested to be hired

After using Talentboost:

• Each candidate takes hardly 2–5 mins for us (We add them into system, send invite, and some cases follow-up)
• Can manage 50–80 Interviews in a day
• Most important, Now we can review a bigger talentpool to get a better fit, typically 100–150 AI screened interviews then 10–15 manual interviews to decide on a hire
• Much better talent due to multi level filtration, We invest time on the most promising candidates now!

Our Take for Future

While unpredictable, we believe tool’s and companies like Althire AI are the way for emerging HR Tech, and anyone not following/using these cutting edge solutions will slowly die-down from this race, Time and time again history has told us, those who do not “evolve” will slowly but steadily be “left out”!

While Althire is one such tool, there are thousand others which are creating ripples across HRTech, In a bigger landscape, We believe that anything that makes life easier and automates the menial work is something that should be invested in!

End Regards

Thank you for following up, If you liked the article, please share with your fellow network and follow me to get notified for more such blogs! If you are interested in what Neurotask does or have any requirements for AI based solution to your problem, please reach out to hsk@neurotaskai.com

If Althire AI intrigues you, here are some of the links:

• Landing Page — https://althire.ai
• Mail — contact@althire.ai
• Linkedin — https://www.linkedin.com/company/althire-ai
• TalentBoost Dashboard — https://talentboost.althire.ai

Uhmmm So What’s the Solution?

Interesting question, but guess what? We heard Azure gives 100$ per year to students under their Student Program plans! Can we exploit this somehow? Obviously yea :)!!!

You make it sound so easy, but how can we as non-technical people do this?

Let me present to you, Azpn, an acronym from Azure Free VPN, This project is a set of jupyter python notebook files which automate the project of deploying a custom VPN server!

Project link — https://github.com/hari01584/Azpn-Azure-Free-VPN

But Why Use This Instead of Other Free VPNs?

Plenty of reasons actually!

• Free VPN are free just on surface, while they steal your data/information in background.
• Unlimited data, you won’t have to worry about bandwidth or any kind of limits!
• Enterprise grade firewall blocks common VPN (example fortinet), Azpn would work well with many kinds of such restrictions.
• The speed is really good! No free public VPN can provide you browsing speed any better than hosting your own on a very robust Azure infrastructure.
• Every command you run is written in jupyter notebook. If you, like my other friends are skeptical about this project, then you can ask your coding/geek friend to verify everything being run in jupyter cells!

Okay, Stop this, I got it, How to install it then?

Install/Deploy

Not much difficult, very easy! Run these commands and follow the steps to get it working :)

1. Go to this azure link to sign-up for student plan, (Done easily if you have @.edu account for verification)

https://www.youtube.com/watch?v=XLxqbEsJ7_I

1. Use the google collab notebook to execute commands for setting up your VPN, here’s the link for google collab notebook Please follow the order of execution of notebooks! Run 1st, then 2nd then 3rd index notebook for smooth Installation process!
2. In your notebook, execute all the cells in order! It will ask for login on this specific cell, where a new window will open and you will need to input a code and login to azure cli (See step 1)

4. Once completed, execute all the cells in order until you get an option to select your region/location of VPN! Write the preferred location(Name) from the list you will get there!

5. Execute all the remaining cells, some cells might take more time than others, for example this one(image below) which takes some time to create and initialize virtual machine! Wait for them to complete before going to next step!

6. After executing all the cells, when you reach Success section, You will find a file named client.ovpn, this is the openvpn config file needed to connect with the VPN, download latest openvpn client and enjoy the fast, secure Internet!

BONUS TIP: You can generate more than one profile in the last bonus section, just enter the number of VPN profiles you want(ie you can distribute the profiles to your friends), and It will generate client1.ovpn, client2.ovpn ….. files that can be used to connect with VPN and share it to others!!

Performance?

Okay, but what are the speed/ping when using it? Can’t you give us the benchmarks?? Yes I can, and would love to do!

Clean Up/Removing Azpn

What if you ever get bored, don’t wana use Azpn, or due to any goddamn reasons decided to remove this service? then What?? Fret not, I got chu covered! Using the next notebook in the series named azpn_remove_all.ipynb, You can remove every changes made by Azpn, and revert yourself to your before-installation state! But why the hell would you do that? :0

Ending Notes/Word of Caution

Since the VPN is hosted in Azure, it means you have full control of it, but since everything is routed through Azure servers then I would like to advise not to use this in illegal activities!

Ending Credits

Kudos to these people for Inspiration!
Raghav Sharma — For telling me how he was hosting his personal VPN on google cloud using free 300$ credits, and then disclosing the trick of using 443 TCP to bypass common VPN detection methods in firewalls!
Rishab Gupta — For having a problem using SonyLiv to watch some webseries using University Internet, being forced to get a free VPN based on Indian servers to access the online content.

Self Hosted Free VPN Using Azure Student Plan with (@.edu) Accounts. [Unblock Fortinet] was originally published in Google Developer Student Clubs TIET on Medium, where people are continuing the conversation by highlighting and responding to this story.