LinkedIn: Leveraging machine learning to solve the modern talent acquisition challenge

Sending an email via Outlook and want to check the recipient’s alma mater? Check. Editing your resume and want to know how others have described the role? Check. Applying for a job and want to get a sense of what your match percentage is? Check. LinkedIn has got you covered.

Now more than ever companies are rethinking their talent acquisition strategies as a changing workforce meets enabling technologies [1]. With that said, the US market in particular is ripe for leveraging job matching services as jobs have become more transactional in nature. In a recent Deloitte study, millennials hold the largest share of the US labor market and 66% of surveyed millennials said they would leave their jobs by 2020 [2]. The migrant nature of modern talent poses challenges for companies as they seek to replace and hire new talent.


LinkedIn’s core value proposition is to “connect talent with opportunity at massive scale” [3]. To effectively connect talent, LinkedIn must build out an intricate network that maps out companies, jobs, network members, schools, and skills. At the core, LinkedIn’s use of machine learning to make these connections is what has allowed the company to be successful thus far and it’s the key to their success going forward. Through their use of machine learning, they can streamline the complex talent acquisition process for talent-seeking companies and improve the job-seeking process for candidates – thus tackling the modern talent acquisition challenge.


LinkedIn has been able to successfully leverage their capabilities and this led them to their IPO in 2011 and acquisition by Microsoft in 2016 [4] [5]. With the latter, LinkedIn now has the potential to integrate itself within the platforms that Microsoft owns. By combining LinkedIn’s machine learning capabilities with Microsoft’s platforms, LinkedIn can continue to develop products that will further connect talent at an even broader scale.


In the short and middle-term and post-integration, LinkedIn has focused on fully integrating its capabilities with the Microsoft platforms. An example of this can be seen with the LinkedIn “resume assistant” that pops up in Microsoft Word as a candidate edits his or her resume [6]. This assistant uses LinkedIn’s network to provide resume wording suggestions and offers potential jobs based on the resume’s content. In the short-term, this business-to-customer approach will enable LinkedIn to continue enhancing their algorithm because they’ll be able to collect even more data as customers/professionals use LinkedIn across Microsoft’s platforms (e.g. connecting Outlook directly to LinkedIn profiles).


For talent-seeking companies, LinkedIn uses its in-house talent-search system called LinkedIn Recruiter to help companies find optimal talent for roles [7]. In the matching process, however, outputs of machine learning can result in biases that may disadvantage already disadvantaged groups. The firm has made an effort to build a representative talent search by adjusting the architecture of the technology to avoid unintended biases. While this gender-neutral approach has been developed, removing biases is still an undergoing effort.


In some ways, LinkedIn is providing a public service as they aim to connect talent across different countries and industries. In the short-term, LinkedIn should ensure that as they integrate with the Microsoft platforms, they devote enough resources to data privacy controls because once this is breached, they will have to deal with the effects of public distrust. As they continue growing, they should also think about segmenting their business by industry and function. Right now, a job-seeker relies on the best match algorithm to find a suitable role but what if the job-seeker doesn’t want a role that’s tied to his or her previous experience? Going forward, LinkedIn should break down the world of possibilities to make the process more digestible for job-seekers.


In the medium to long-term, LinkedIn may have to pivot to a more business-to-business model because talent-seeking companies will want this data synthesized in a way that enhances their talent acquisition strategies. In addition, as LinkedIn continues to build out its capabilities, they will have to be cognizant of emerging competitors. There are already several HR tech companies (e.g. Mya, Ideal, Entelo, Restless Bandit) using artificial intelligence to connect talent with opportunities [8]. These competitors will not have a robust set of network data like LinkedIn does but given their expertise, they can synthesize and customize what they do have for talent-seeking companies.


LinkedIn is well positioned to accomplish its mission of connecting talent on a massive scale. Its technology, however, leads to several questions. As society moves to embrace diversity (ethnic, sexual orientation, etc.), how will LinkedIn’s machine learning incorporate these aspects into its algorithm? What legal implications will this uncover? LinkedIn will have to be cautious and ensure equal and fair representation – tasks not easily auditable by artificial intelligence.


Additionally, how will talent management and HR teams at firms react to LinkedIn’s advanced capabilities? Will they push against it and drive forward their own proprietary systems or maintain old-fashioned approaches? Will they pull in LinkedIn and fully rely on their technology to determine the future of their firm’s talent? Whatever approach firms choose will ultimately determine their livelihood.


(789 Words)




[1] Harvard Business Review, “The Future of Talent Acquisition,”, accessed November 2018.

[2] Deloitte, “The 2016 Deloitte Millennial Survey Winning over the next generation of leaders,”, accessed November 2018.

[3] Jeff Weiner, “From Vision to Values: The Importance of Defining Your Core,”, accessed November 2018.

[4] LinkedIn, “A Brief History of LinkedIn,”, accessed November 2018.

[5] Rick Gillis, “Microsoft’s Acquisition Of LinkedIn Changed The Job Search Industry In Ways We Don’t Even Know Yet,”, accessed November 2018.

[6] LinkedIn, “LinkedIn Talent Solutions,”, accessed November 2018.

[7] Sahin Cem Geyik and Krishnaram Kenthapadi, “Building Representative Talent Search at LinkedIn,”, accessed November 2018.

[8] Ascendify, “How AI is Disrupting (and Enhancing) Talent Acquisition,”, accessed November 2018.

[Cover Image], accessed November 2018.


The Rise of Automotive Additive Manufacturing


Applications of Machine Learning in the US Criminal Justice System

Student comments on LinkedIn: Leveraging machine learning to solve the modern talent acquisition challenge

  1. This is a very timely issue and one that isn’t unique to LinkedIn. With GDPR, a lot of machine learning services are being asked to “audit” the decisions their algorithms make. It will be very interesting to see how different societies deal with trade-off of a faster pace of machine learning progress with the ability to make sure those machine learning decisions are fair and equitable.

  2. I agree that LinkedIn may want to start competing with the other talent acquisition players, but I wonder if doing so without careful thought would detract from user engagement. It’s important that LinkedIn be useful to non-jobhunters and jobhunters alike to maintain overall engagement on the platform. I think if LinkedIn goes B2B, it will need to do other things in tandem that enhance the B2C value prop to ensure the network stays strong.

  3. I think the answer to your very last question depends on whether a company has an access to the targeted talent pool they are looking for. For instance, consulting firms would not have incentives to use LinkedIn to recruit a general consultant, as they already have many candidates and they have established database on their own. However, many consulting firms are now looking for a new type of customers (e.g., digital, design) in which case the LinkedIn services would be very helpful.

  4. I believe the answer to the last question of yours comes down to talent in recruiting. With the massive inequality in the demand and supply of data scientists, the winner of this game will be the one that is able to recruit and retain resources. Being under the large entity of Microsoft, LinkedIn right now has an advantage of attracting talent, and as more and more firms start moving towards a data-driven recruitment platform, I see this as the biggest challenge they must overcome for developing the best capability.

  5. Very astute writing! I’m left with a more fundamental question of how much we want to use machines to screen candidates for employers. At the same time that those machines may be less susceptible to bias (assuming they can be programmed in a way that overcomes the bias of the data that they are fed), to what extent are machines able to screen better for human potential or how well different people can work together or cultural fit? In a time where Harvard, in particular, is being challenged for perceived bias in its admissions process, is the way to go a more AI-centered approach?

    Also, the acquisition by Microsoft seems promising for LinkedIn’s ability to maintain a competitive advantage in this space; with it’s ability to provide LI with the scale and technology brainpower to build an AI platform.

Leave a comment